The following is the conclusion of a two-part series. The first part was posted on Wednesday, March 14.

Was VNS wrong in its projections?

Voter News Service (VNS)—the organization that projects the outcome of races on election day for the five major American television networks and Associated Press (AP), based on vote tabulations and exit polls—has attempted to explain its “mistakes” in Florida in a report to its members. At 7:50 p.m. on November 7, VNS projected Democratic candidate Al Gore a winner by 7.3 percent over Texas Governor George W. Bush. The VNS report, as cited by CBS and CNN, mentions only percentages, not actual numbers. But based on a voter turnout marginally smaller than the 5.96 million who actually cast ballots in Florida on election day—VNS somewhat underestimated the voter turnout in Florida—its projection of a 7.3 percent edge would translate into a margin for Gore of some 400,000 votes.

VNS attempted to account for this 7.3 percent error by breaking it down into four parts: (1) Errors in exit polls, 2.6 percent; (2) Use of the 1998 governor's race to construct computer models, instead of the 1996 presidential or 1998 senatorial contests, another 2.6 percent; (3) Mistaken estimation of the absentee vote, 1.7 percent; (4) Quality control problems, 0.4 percent.

Assuming the erroneously projected margin for Gore to total 400,000 votes, then errors in exit polls would account for 144,000 votes, the use of the wrong election model would account for another 144,000 votes, the mistake on absentee ballots would account for 92,000 votes, and quality control problems, the remaining 20,000.

(A 40,000-vote mistake in Gore's favor in Duval County apparently made by a keypunch operator, which has been referred to in the media, did not take place until an hour after the call had been made for the Democratic candidate [the error occurred at 9:07 p.m. and was corrected by 9:38] and thus had no impact on the 7:50 p.m. call.)

One anomaly in the VNS analysis immediately presents itself. The 1998 gubernatorial election, used as a model by VNS, was a relatively comfortable victory for Republican Jeb Bush over his Democratic opponent, by 55.3 to 44.7 percent. The 1996 presidential and 1998 senatorial races—which VNS officials suggest would have provided a more accurate picture of voting in 2000—were both won by Democrats (Clinton, by 5.7 percent, as mentioned above, and Bob Graham in the Senate race, by a landslide margin of 25 percent, respectively). On the face of it, it seems strange that a model based on a decisive Republican victory would more likely to lead to a call for Gore than one based on a Democratic victory.

The mistake in the estimate of absentee voting is sizable. VNS projected that 7.2 percent of Florida voters would be absentee voters, when, in fact, the latter represented 12 percent of the total (approximately 700,000 people). Moreover, VNS assumed that absentees would be 22.4 percent more Republican than election day voters. They turned out to be 23.7 percent more Republican.

Any discussion of absentee voting, which is increasing nationally as a percentage of the total vote, is obliged to take into consideration the well-known fact that this form of voting is particularly vulnerable to abuse. A November 14, 2000 article in the St. Petersburg Times (“Florida's Vague Absentee Laws Make Fraud a Concern”) noted that absentee voters are required to sign an affidavit “stating they are unable to go to the polls because they are disabled, will be out of town on Election Day, have recently moved or have moved to a state in which they can't vote.”

In practice, this is very loosely enforced. In the case of out-of-state voters who still claim Florida residence, according to this article, “Florida laws governing absentee ballots are so vague that virtually anyone who declares the Sunshine State their home can get an absentee ballot, as long as they claim to be temporarily away. Even if temporary is 15 years.”

The article continued: “‘Numerous' questionable absentee ballots were cast [on election day 2000] from Bay County, a GOP stronghold in Florida's Panhandle where Republicans had an aggressive effort to gather absentee votes, according to an affidavit by a Panama City woman. Prosecutors in Escambia County are investigating at least one absentee ballot that was stolen from a Miami man and forged...

“Several Tampa Bay area voters say they suspected something was amiss when they were turned away from polling places after being told—erroneously—that they had punched absentee ballots. In Hillsborough, scores of voters received telephone calls before the election from people claiming to work for Hillsborough Elections Supervisor Pam Iorio. The callers, who claimed to work for the ‘elections commissioner,' falsely told voters who had requested absentee ballots that the ballots had not been received.”

In Seminole and Martin counties Republican election officials were allowed to alter thousands of absentee ballot requests—to fix a printer's error—several weeks before the election.

Moreover, the results of the 1997 Miami mayoral election were thrown out when a court found that the staff of the apparent victor, Republican Xavier Suarez, had tampered with as many as 4,740 absentee ballots. His opponent, Joe Carollo, was installed in the office. Absentee ballots were shown in court to have been forged, coerced, stolen from mailboxes or fraudulently obtained.

Suarez to this day sits on the executive committee of the Miami-Dade Republican Party. Following last November's election, he told a journalist, “Dade County Republicans have a very specific expertise in getting out absentee ballots. I obviously have specific experience in this myself.”

Gap between exit polls and vote returns

The most politically significant aspect of VNS's review of its supposedly mistaken call for Gore concerns the gap between the exit poll findings and the actual vote. Given VNS's track record, such a large discrepancy is extraordinary.

Ted Savaglio, a leading VNS official, told the Congressional panel on February 14, “Since 1990, when the first joint polling and projection effort began [the creation of Voter Research and Surveys (VRS), one of the forerunners of VNS], we have been involved in nearly 900 election contests across the nation. The methods that we used to project winners in those races have only been wrong once before. In other words, we have been right 99.8 percent of the time.”

In its internal report, VNS officials, presumably referring to the use of models before the existence of VNS or VRS, asserted that “these models have been used to call approximately 2,000 races, and only six errors have occurred.” The CNN report comments, “The odds are more like 1 in a million for a miss of 7.2 [sic] percentage points.”

The CBS report makes much of the fact that VNS made four retractions November 7 (counting the calls for Gore and George W. Bush in Florida each as one). The other two involved projecting (and later withdrawing) a victory for Gore in New Mexico and for the Democratic candidate for the US Senate, Maria Cantwell, in the state of Washington.

It should be noted, however, that these calls turned out to be not mistaken, but merely premature. Gore did, in fact, end up winning the vote in New Mexico (established by a recount that took several weeks), as did Cantwell in Washington.

Moreover, VNS waited to call New Mexico for Gore until nearly an hour-and-a-half after the polls had closed and with about half the precincts included in the count. In Washington, its analysts waited for nearly two hours before calling the race for Cantwell. In other words, neither case was similar to the situation in Florida, where, within 40 minutes, VNS was quite confident of its projection, a victory for Gore by a margin of hundreds of thousands of votes.

How is the enormous gap between the exit poll numbers and the actual vote totals to be explained? One way to answer (or, more precisely, dodge) the question is to place all the blame on VNS. This is largely what has happened.

Fox has threatened to pull out of VNS and the other television networks have demanded major changes in its organization and methods. The networks argue that VNS seriously needs repair (or replacement), that its models and computers are out of date, that it needs to improve its methodology, that it needs to hire more staff and spend more money, etc. This may very well be true. However, much of this is a smokescreen thrown up to preempt an investigation into the large discrepancy between what voters told VNS pollsters as they were leaving polling places, and the vote totals subsequently reported by Florida election officials.

The only explanation offered by VNS for the discrepancy between its exit poll projections and the actual vote totals in Florida is that its exit poll sample was too small, and therefore prone to error. VNS does not, however, indicate that its exit poll sample in Florida was smaller, in percentage terms, than that used in other states.

Unless one assumes that in Florida on November 7, 2000, as opposed to some 900 previous races, VNS pollsters and statisticians inexplicably came down with a catastrophic case of incompetence, the explanations provided by both the networks and VNS itself do not appear to hold water.

There is another possible explanation for the gap between the exit polls and the official vote totals that is at least as plausible as those offered by VNS and the networks.

The CBS report hints at this, when it casually observes: “In the Florida exit polls, people reported how they had voted, assuming their votes were being counted. Some may not have been.” Later it states: “The exit-poll sample estimated a significant Gore lead that never materialized.” And once more the report notes obliquely: “[A]s we have seen in the Florida recount, what voters think they have done at the polling place may not be reflected in the totals when the votes are counted.”

People assumed their votes were being counted, and they may not have been. Why not? Because of machine error, a confusing ballot, bureaucratic indifference and neglect, outright fraud?

There is the distinct possibility that the VNS projections were not so far off after all, and that the votes of tens of thousands of people, for a variety of reasons, were not included in the totals. This side of the matter, which involves the basic democratic rights of the population, particularly its more oppressed layers, barely causes the CBS authors to pause and does not even occur to the authors of the CNN report. Or at least they choose not to comment on it. They may have felt that this would “risk placing ... an extra twig” on the fire.

At the February 14 hearing in Congress, Rep. Peter Deutsch, Democrat from Florida, raised the possibility that the networks' exit polls had been right and that “if the vote had been counted in Florida, Al Gore would be president of the United States.”

VNS officials are not speaking to the press about this or any other matter. When asked by a WSWS reporter about the possibility that its projections had been correct and that “other processes had intervened,” a VNS spokeswoman immediately referred to Deutsch's comment and conceded that such a course of events was at least “possible.”

We know that thousands of Gore votes were discarded, as overvotes or undervotes, some due to confusing ballots, as many as 10,000 in Palm Beach County alone. One-fifth of the votes in black communities in Duval County (Jacksonville) were thrown out. The existence of punch-card ballots in poorer neighborhoods ensures that working class voters have a more difficult time having their votes counted.

Over and above these factors, there is another possibility, one that is excluded in advance by the networks and all other official institutions. This possibility, however, would be the subject of intensive investigation by anyone honestly and conscientiously interested in getting at the truth—namely, that thousands of votes were deliberately misplaced, altered or otherwise disposed of.

There is no reason, other than political expediency, to rule out the possibility of widespread fraud, vote tampering or other forms of criminal activity on the part of the Bush camp, including Florida Governor Jeb Bush and Secretary of State Katherine Harris.

Did the Bush camp originate the call for Bush?

The CBS report is obliged to acknowledge that “the first Florida call for Gore was probably unavoidable, given the current system of projecting winners.... However, the second Florida call, the one for Bush, could have been avoided.”

Given all the circumstances, including the retracted call for Gore earlier the same evening, the call for Bush, even taking into account VNS errors, is nearly inexplicable except as the result of chicanery or wishful thinking.

Between 10 p.m. and 2 a.m. Bush held a lead, but not one decisive enough to warrant a call. Shortly after 2 a.m., a 20,000 vote change in Volusia County, which was later claimed to be the result of a faulty computer memory card, increased the Texas governor's VNS-calculated statewide lead to more than 51,000 votes. The VNS determined that there were only 180,000 votes remaining to be counted. It turned out that, in fact, there were 400,000 outstanding votes and Bush's lead was less than 30,000, with some traditionally Democratic counties (Dade, Palm Beach and Broward) remaining to be heard from.

At about 2:10 a.m. the network decision teams, with Fox (and Bush cousin, John Ellis) at the head of the pack, began to contemplate a call for Bush. Remarkably, they all claim that they failed to cross-check VNS totals with those of AP, which showed Bush with a smaller lead and losing ground. Again, the networks blame their premature call for Bush on their reliance on VNS as the sole source of information.

But it is a fact that neither VNS nor AP ever called Florida for Bush. That decision was made by the networks acting on their own.

At the time Fox called Florida—and hence the national election—for Bush, at 2:16 a.m., AP (as a result of correcting the Volusia mistake) was about to report that Bush's lead had dropped to 17,000 votes. NBC, CBS and ABC followed Fox's lead within four minutes.

The facts raise the distinct possibility that (1) Ellis acted in consultation with the Bush forces when he decided to call Florida for Bush; and (2) he did so at 2:16 a.m. not because the data at that time was confirming a Bush victory, but for precisely the opposite reason: the information indicated that Bush's margin was fading and might very well disappear.

According to this hypothesis, Fox, acting as the media arm of the Bush campaign, via Bush family member Ellis, moved to preempt the possibility of an eventual call for Gore by calling Florida for Bush—prematurely and knowingly so—with the aim of stampeding the other networks into following suit and cowing Gore into making an early concession. If this was the scheme, it succeeded in the first aim, and came very close to succeeding in the second.

Such manipulation would constitute a massive and criminal act of electoral fraud—a de facto political coup. But it is certainly in keeping with the known facts. A rejection, out of hand, of this possibility on the grounds that the Bush camp would never descend to such methods could only be the product of political amnesia or dishonesty. Involved here, after all, are the same political party and the same media establishment that paralyzed the federal government for more than a year in an attempt to topple the Clinton administration by means of dirty tricks, witch-hunting, entrapment and similar methods. Moreover, the post-election day campaign of Bush to block a recount of Florida votes could only increase, in the mind of an objective observer, suspicions about the actions of the Bush camp on election day itself.

Questions also suggest themselves about the 20,000 vote “mistake” in Volusia County, which proved to be so convenient for the Bush camp. It apparently had a significant impact on the networks' calling of the Florida and national vote for Bush. This would be one more fruitful area of investigation in a serious probe of the national election.

Incidentally, the argument raised in some quarters that the other television networks were about to call Florida for Bush at around 2:15 a.m. is hardly a vindication of Ellis's role. It merely points to the sinister role played by all the networks on November 7-8. Any serious analysis at that time of the data, or a simple check of the AP's figures or the Florida Secretary of State's numbers, would have indicated that Bush's lead was questionable and that it was, in fact, likely to vanish.

A rumor about GE Chairman Jack Welch

It is apparently common knowledge within media and government circles that certain events on election night point to something more culpable than inadvertent error on the part of the networks. At the February 14 Energy and Commerce Committee hearing, according to a report in the Washington Post, Rep. Henry Waxman, a Democrat from California, “asked about what he termed a ‘rumor' that Jack Welch, chairman of General Electric, NBC's corporate parent, ‘intervened to call the election for George Bush.' He offered no proof. NBC's [news division president Andrew] Lack called that ‘untrue,' ‘rather foolish' and ‘just a dopey rumor,' saying Welch was watching the coverage in the newsroom but not calling the shots. Waxman said the panel might subpoena an internal tape of Welch made that night.”

Lack's response is significant. He concedes that the Welch story is circulating and that the corporate boss, one of the most powerful figures in American business, was seated in NBC's newsroom. Welch's readiness to intervene at the network is apparently legendary.

The very fact that NBC News allowed the head of its corporate parent to sit in its newsroom and monitor its election day coverage is a damning indictment of the network. It makes a mockery of the supposed impartiality of the network's news coverage. The fact that no media outlet has raised so much as a peep about this revelation is indicative of the corrupt and compromised state of the entire media establishment. How many other corporate bosses at other networks “sat in” while their high-paid employees spun the election day news to the public?

As for Welch's ties to Bush, the Dallas Morning News reported January 4, 2001 that Bush had hosted a meeting with executives from “General Electric, Eastman Kodak, Cisco Systems, Boeing, Enron, Wal-Mart and General Motors. Many were contributors to Mr. Bush's campaign ... GE chairman Jack Welch said the ‘fiscal stimulus' of the Bush tax cut would complement the monetary benefit of the interest rate cut. ‘We raised a glass of water to Mr. Greenspan for taking that action,' Mr. Welch said after what he described as a positive meeting with the president-elect.”

There is no indication that Congressman Waxman has taken any steps to subpoena the tape of Welch on election night. A call to Waxman's office elicited no response.

At approximately 2:30 a.m. on November 8, Al Gore telephoned George W. Bush and conceded the election. Fifteen minutes later AP reported that Bush's lead was down to 13,934, while VNS was still claiming he had a 55,449 vote lead. At 2:51 VNS apparently discovered its Volusia County error and by 2:55 announced a lead of only 9,163 for Bush. Some time in the next half hour Gore was reached in his limousine, en route to making a concession speech, and alerted to the actual vote total. He subsequently phoned Bush and retracted his concession.

By 3:40 a.m. the Bush lead had dropped to 6,060 votes. Between 3:57 and 4:05 a.m., with Fox being the last, the networks withdrew their calls of a Bush victory. At 4:10 Bush's lead had fallen to some 1,800 votes, where it would remain until the recounts began.

A farcical post-script: the House committee hearing

The February 14 appearance before a Congressional committee by all the television news chiefs, the president of AP, the authors of the CNN report and the outside consultants on the CBS and NBC studies was both an anti-climax and a charade.

Bush's selection as president December 12 by the majority on the Supreme Court, and the fact that Rep. Tauzin's charges of anti-Bush bias leveled against the networks were easily disproved, had served to cool the Louisiana Republican's ardor. What precisely was the purpose of the hearing, once the accusation of political favoritism had been more or less set aside? No one seemed clear.

It was still billed as an inquiry into “Election Night 2000 Coverage by the Networks,” but Tauzin and the others on the panel readily professed they had no right telling the networks how to cover the election. In the event, the television news chiefs dutifully appeared before the committee, complained about having to show up and proceeded to repeat the mea culpas they had been uttering since November 9. “Our Florida flip-flops are deeply embarrassing to us,” said CBS News President Andrew Heyward. “We let our viewers down,” declared Fox News President and former Republican Party operative Roger Ailes. “Make no mistake about it, we are embarrassed by those errors,” commented NBC's Lack.

Lack and AP's Louis D. Boccardi both made reference to the problems experienced by many Americans November 7 in voting and having their votes counted. According to the New York Times, Lack “said the mess that unfolded after election night was a revelation, in that it was clear that being registered to vote did not necessarily mean that one could vote, and that people who are poor have a harder time voting than people who are rich.”

In his statement, Boccardi asserted “that if voters were actually discouraged by media projections from casting ballots on November 7—and we have seen no credible evidence to show that many were—their number is eclipsed by the tens of thousands of voters in Florida and millions nationwide who were disenfranchised by voting machine breakdowns, confusing ballots, lost votes, and a host of other consequences of official error, disorganization and incompetence in administration of elections.”

Not wanting to let go entirely of his “hot button” issue, Tauzin claimed that the networks' exit poll model contained “statistical biases in favor of Democrats,” although he now admitted that there was “no evidence of intentional bias.” Rep. Edward Markey of Massachusetts, a Democrat, termed “preposterous” these suggestions of a “vast left-wing conspiracy.”

For their part, a number of Democrats pontificated and played to the crowd, raising issues that they hadn't previously pursued and had no intention of pursuing following the hearing. In addition to repeating the rumor about GE's Welch having called the election, Waxman criticized Fox for employing Ellis, Bush's cousin. He asserted that Ellis's projection of a Bush victory “created a presumption that George Bush won the election and set in motion a chain of events that were devastating to Al Gore's chances.”

Ohio Democrat Sherrod Brown asked, in the light of Fox's use of Ellis, whether the networks should have a policy against hiring candidates' relatives. Ben Wattenberg, one of the authors of the CNN report, responded, “I think that's preposterous.” Brown replied by calling Fox “the most conservative politically of the major networks.” Wattenberg, a fellow at the right-wing American Enterprise Institute, thereupon asserted, “There are a lot of Americans who think the other networks are too liberal.” Rep. Bobby Rush, an Illinois Democrat, demanded a hearing on minority voters being “harassed by police across the country.”

In reality, the Democrats were, if anything, less interested in investigating the events of election night 2000 than the Republicans. If the Democrats on the committee had been genuinely concerned with looking into the role of the television networks, they would have subpoenaed records and telephone logs, or at least publicly demanded that the committee take such action, and insisted on calling many more witnesses, including all the network news anchors and analysts, John Ellis and all the other decision team members at the various networks, and the principal political actors, including George W. Bush, Jeb Bush and Katherine Harris, as well as top Bush aides like Karl Rove and Karen Hughes. They would not have missed the chance to call Jack Welch. None of that took place.

The lethargic attitude of the Democrats toward the hijacking of the presidential election contrasts sharply with the attack-dog aggressiveness of the Republicans when it comes to sensational probes of Clinton and the ex-president's associates. Republican Congressman Dan Burton routinely subpoenas everyone and everything in sight, not on the basis of evidence of criminal activity, but rather on the grounds of suspicions (his suspicions) that the possibility exists of wrong-doing.

The February 14 House hearing, as much as anything else, demonstrated the complicity of the Democrats in the attack on democratic rights at the center of the 2000 election crisis.

We now know enough about the night of November 7-8, 2000 to insist that the events surrounding the election warrant a serious investigation. Such an inquiry would have to be carried out on several fronts.

In the first place, George W. Bush and Florida Governor Jeb Bush would be obliged to tell, under oath, the content of their discussions concerning the Florida vote. Why was Jeb Bush so confident about the outcome of the Florida vote? Did he know something the general public did not?

In a serious investigation, Jeb Bush, Katherine Harris and other state Republicans would be questioned, under oath and at a public forum, about the conduct of the vote itself. The role of individuals such as Xavier Suarez and other Florida Republicans would be probed, in regard not simply to the absentee vote, but to the actual tabulation of votes, with an eye to possible vote tampering.

According to CBS, Miami and Tampa “had the biggest overstatement of the Gore lead in the exit polls.” An investigation might begin with these two cities, particularly in light of Miami's recent experience with a corrupt Republican Party machine.

The 20,000 vote error in Volusia County needs to be looked into. Which individuals were responsible for this supposed mistake?

The conduct of the television networks and their decision teams, including John Ellis, needs to be thoroughly examined. Did the networks rescind their call for Gore at 10 p.m. November 7 in response to pressure from the Bush camp?

What exactly did Ellis discuss with the Bush brothers? What role did they and their aides—or, for that matter, their father, the former president—have in Ellis's decision to prematurely call Florida for Bush in the early morning hours of November 8?

Why did the television decision teams fail to cross-check their VNS vote totals with those from AP and the Florida Secretary of State, available on the Internet? What was the role of Jack Welch of GE and other corporate executives in the decision-making process?

Given the facts surrounding the vote—including VNS's projections of a sizable Gore majority, the crucial character of the contest, the role of Jeb Bush as head of the Florida state apparatus—and the stakes—a national election—the circumstances of the Florida vote warrant both a public probe and a criminal investigation.

A further lesson

In late February, the television networks and mass media, after promising in public to reform and never again jump to precipitous conclusions, demonstrated that something more than a momentary aberration had been at work in November. When the Miami Herald on February 26 revealed the results of its recount of Miami-Dade's “undervotes” (ballots that didn't register a presidential preference when counted by machine)—a gain by Gore of some 49 votes—the news media leaped at the chance to announce that “Gore Still Loses.” The headlines were misleading, as usual.

The Herald study examined undervotes in Miami-Dade and added the results to the totals reached in manual recounts carried out during the post-election crisis by local election boards in three other counties—Broward, Palm Beach and Volusia. These are the four counties in which Democrats had called for hand recounts.

In making the claim that the results of its Miami-Dade recount showed that Bush “really won” in Florida, the Herald and its media associates simply ignored other partial recounts that point in the opposite direction. In any event, all indications are that a full statewide recount of undervotes, as ordered by the Florida Supreme Court and countermanded by the US Supreme Court, would result in Gore gaining thousands of votes.

More fundamentally—and this is an issue which the media refuses to discuss—whether or not a recount in the end might have awarded Bush the greater number of officially recognized votes, his manner of attaining office, through the suppression of votes and an attack on the principle of popular sovereignty, has eternally branded his administration as both antidemocratic and illegitimate.

As the foregoing analysis of the networks' whitewash of their own role in the events of election day 2000 has shown, these corporate-controlled institutions are deeply implicated in the assault on the democratic rights of the American people which was at the heart of the 2000 election crisis—an attack that did not end, but rather entered a new stage with the installation of the Bush administration.