Following the 2000 election a popular bumper sticker in Florida read, “If you think we can’t vote, wait ‘til you see us drive.” But nobody’s laughing now.
Unless the Democratic-controlled House votes to hold a new election, Republican Vern Buchanan will represent Florida’s 13th congressional district in the 110th Congress. Buchanan “won” election by beating Democrat Christine Jennings by 369 votes out of 237, 861 cast.
The election was held on direct-recording electronic (DRE) voting machines that don’t leave a paper trail. Strangely, according to the machine count, 18,382 voters in Democratic Sarasota County did not bother to cast a ballot in the congressional race (that’s 15 percent of those who voted in the county on November 7). That is unlikely, since it was a hotly contested battle for the seat vacated by former Secretary of State Katherine Harris, who ran for Senate and was soundly defeated. (Yes, Al, there is a God.)
All evidence points to voting machine malfunction. If those 18,382 votes had gone to Jennings by the same percentage with which she carried Sarasota County, 53 percent, she would have won the election by 600 votes.
The machines in question are iVotronic DRE touch-screens manufactured by Election Systems & Software (ES&S). According to a Sarasota Herald-Tribune survey of election workers, one in three reported that they had received complaints from voters who said their votes did not register. And more than 100 voters have told Jennings’ campaign that, after voting on the iVotronic touch screen machines, their votes for Jennings did not show up on the summary displayed at the end of the voting process.
A recount was conducted, but little changed, which is only natural. One can’t really recount votes cast on DRE voting machines – one can only re-run the vote-counting program that caused the problem in the first place. Nevertheless, Gov. Jeb Bush, a champion of electronic voting, certified the recount and promised that the Sarasota machines would be audited – by Alec Yasinsac, a Florida State University professor who during the 2000 election imbroglio held vigil on the steps of the Florida Supreme Court wearing a “Bush Won” button.
It is up to the Democratic leadership in the House to decide whether or not to call for a re-vote. MoveOn.org has organized a national petition drive aimed at Speaker Nancy Pelosi. Short and sweet, the petition reads, “In the wake of Florida’s electronic voting machine meltdown, Congress should call for a re-vote and repair our nation’s elections.”
Florida Republicans, skilled in the art of stealing elections, have geared up for battle. Miami Rep. Mario Diaz-Balart told the St. Petersburg Times, “If that happens … forget about any possibility of pretending to be civil. That would be an affront to democracy.”
The 2006 Exit Poll Discrepancy
With a couple of notable exceptions the 2006 midterm elections were not fraught with the problems that were endemic in 2004.
Unlike 2004, in 2006 there was no glitch in the CNN computer system, so we don’t have access to the unadjusted results of the exit poll that was conducted for ABC, CBS, CNN, NBC, Fox and AP. And unlike 2004, the two firms conducting the exit polls, Edison Media Research and Mitofsky International, waited until 5 p.m. EST to release their exit polls to their media clients.
However, after 5 p.m. EST someone at CNN did leak some of the exit poll results for 10 Senate races to ThinkProgress, the blog of the Center for American Progress in Washington.
In two key races for the Senate, the exit polls indicated election results that diverged significantly from the official results.
In Montana, the exit polls indicated that Jon Tester would win 53 percent of the vote to Conrad Burns 46 percent. However in the official count Tester eked out a 49 percent to 48 percent victory over Burns. In other words, there was a 6‑percentage point discrepancy between the exit polls and the official count.
In Virginia, the exit polls indicated that Democrat Jim Webb would beat Republican incumbent George Allen, 52 percent to 47 percent. However, in the official count, Webb and Allen were virtually tied, with Webb getting 49.59 percent of the votes and Allen 49.20 percent of the votes (a difference of 9,329 votes). In other words, there was a 5‑percentage point discrepancy between the exit polls and the official count.
As Steven Freeman and I explain in our book Was the 2004 Presidential Election Stolen? Exit Polls, Election Fraud and the Official Count, in 2004 a similar pattern occurred. The 2004 Election Day exit polls showed Kerry did better than the official count indicated in the key 11 battleground states, but most significantly in Nevada, New Mexico and Ohio.
As in 2004, in 2006 the discrepancy between the exit poll results and the official count in Montana and Virginia raise questions about the integrity of the voting processes in both states.
This is particularly true because, in almost all other cases, the 2006 exit polls were on the money. For example, in the hotly contested Missouri race, the exit polls indicated that Democrat Claire McCaskill would beat the Republican incumbent Jim Talent 50 percent to 48 percent and the official vote total was 50 to 47. And in Tennessee, the exit polls predicted that Harold Ford would lose to Republican Bob Corker 48 to 51 percent, and the official vote total was 48 to 51 percent. These results indicate that the pollsters’ polling methodology was on the mark. So why were the discrepancies in Montana and Virginia so large? Did voting machines malfunction?
If we had access to the precinct-level exit poll data from Montana and Virginia, which the pollsters have withheld, we would be able to investigate whether the size of the exit poll discrepancy correlates with the voting technology used in specific precincts.
— — — — — — —
Take Montana: In the Big Sky State, 16 counties used paper ballots and 40 counties use optical scan voting systems manufactured by ES&S, one of the three largest voting machine companies. If we had access to the exit poll data, we would be able to compare the size of the discrepancy in the precincts in the 16 counties where Montana voters cast paper ballots with those precincts in the 40 counties where voters used ES&S optical scan machines.
The 2004 exit poll data indicated that in rural and small-town precincts – the only precincts for which the exit pollsters released data that could be meaningfully compared – the difference between the exit poll results and the official count was three times greater in precincts where voters used machines than in precincts where voters voted on paper ballots. In other words, in precincts that used paper ballots there was no discrepancy between the exit polls and the official count. It would be interesting – and if we had the data, very possible – to know whether or not that was the case in Montana.
Unlike the voting systems manufactured by competitors Sequoia and Diebold, ES&S optical scan machines have not been publicly hacked. However a study by the Brennen Center for Justice released in October found:
All of the most commonly purchased electronic voting systems [including ES&S optical scan systems] have significant security and reliability vulnerabilities, which pose a real danger to the integrity of national, state, and local elections. … The Brennan Center’s Task Force on Voting System Security reviewed more than 120 potential threats to voting systems. Among its key conclusions was the finding that attacks involving the insertion of software attack programs or other corrupt software are the least difficult attacks against all electronic systems currently being purchased, when the goal is to change the outcome of a close statewide election.
— — — — — — —
In Virginia, the situation is less clear. The 95 counties in Virginia use 17 mechanical voting systems manufactured by seven different voting machine companies, most of which are some form of DRE electronic voting systems that leave no paper trail.
Lest their be any doubt about the dangers of electronic voting, in November the National Institute of Standards and Technology (NIST), a federal agency, issued a draft report that noted:
The computer security community rejects the notion that DREs can be made secure, arguing that their design is inadequate to meet the requirements of voting and that they are vulnerable to large-scale errors and election fraud.
And NIST concluded, “Software dependent approaches such as the DRE are not viable for future voting systems.”
The 5‑percentage point discrepancy between the exit polls and the official count in Virginia indicate that vote fraud is a plausible possibility. Yes Democrat Webb won, but if the election for the Virginia Senate seat was hacked, that would have involved a pre-programmed electronic transfer of votes from one candidate to another – all of which would have been done so as to not to raise suspicion by interfering with the total number of votes cast. That raises the question: Did Webb win because GOP operatives didn’t program the electronic voting systems to steal enough votes?
We can only speculate. This, however, is certain: A thorough examination of the precinct-level exit poll data and the voting technologies used in those precincts where the exit polls were conducted could lay help lay such questions to rest once and for all.
Joel Bleifuss, a former director of the Peace Studies Program at the University of Missouri-Columbia, is the editor & publisher of In These Times, where he has worked since October 1986.