2002 Alabama general election

Jump to: navigation, search

Alabama's 2002 election saw popular Democratic governor Don Siegelman running for reelection. In an incredibly close race, Siegelman initially won the state, but after Baldwin County Republicans adjusted the numbers overnight, his GOP opponent Bob Riley was ahead. Statistical analyses pointed to electronic vote tampering, and Siegelman soon underwent a partisan assault to end his political career, indicating he was a target for the GOP.

Pre-election polls


GOP fixing in Baldwin County

Baldwin County's results changed multiple times on election night, flipping the state's winner. At first, an obscure Libertarian candidate captured thousands of votes, which Siegelman and Riley's teams both recognized as absurd. Election officials fiddled with the central tabulator and produced the first real set of results, which showed Siegelman had 19000 votes. Despite losing the county, this was enough for Siegelman to win the state by a razor-thin 0.2% margin. Siegelman declared victory, and it looked to everyone like he'd won. However, after the poll watchers went home, the GOP election officials who stayed behind conducted another recount and produced a third set of results. They showed Siegelman with only 13000 votes, losing the governor's race.[1][2]

Democrats thought the overnight alteration smelled fishy, and accused the Republican officials of shenanigans. But the Republicans defended the change, pointing out that the first tally giving Siegelman 19000 was impossible because there were more votes than voters. Who was right?

Statistical anomalies

Auburn University political scientist James Gundlach found statistical signs of electronic tampering. He concluded that Baldwin County's results were rigged for Bob Riley, and Siegelman was the true winner of the state.[3]

In terms of a "smell test", Riley's gain over the 1998 Republican candidate was strange. Riley did 10000 votes better than Fob James, the 1998 Republican who came from Baldwin County. Though Riley did better than James statewide, no other county showed such a massive increase for him, not even Riley's own home county.

Gundlach used linear regression to rigorously establish Baldwin County as an outlier. Across the counties, Siegelman's 2002 vote share linearly correlated with his 1998 vote share. However, the second set of Baldwin results (from the overnight GOP recount) moved Baldwin County far away from the linear trend. To test the counterargument that George W. Bush campaigning for Riley might have made the difference, Gundlach also plotted Riley's 2002 vote shares against Bush's 2000 vote shares. Once again, there was a strong linear correlation across the counties, but the second set of Baldwin returns moved it far away from the linear trend. That supposedly fixing an error would make Baldwin County's results more irregular raises suspicion.

He then looked at how Siegelman's vote share changed between 1998 and 2002 within individual precincts. Gundlach compared Montgomery and Shelby County to Baldwin County. In both county groups, Siegelman's 1998 results accurately predicted his 2002 results. But in Montgomery and Shelby County precincts, he got 85% of his 1998 vote (the same as he did statewide), while in Baldwin County, he got 70% of his 1998 vote. This further establishes Baldwin County as an extreme outlier from the rest of the state.

Baldwin County's unexplained deviation from statewide trends indicates that its results were fraudulent. Gundlach calculated that 18% of Siegelman's votes were flipped to Riley in each precinct.

But what about the first set of returns, which had more votes than voters? Gundlach asserted that it was caused by an initial error with flipping the votes. Between the first and second Baldwin County results, Siegelman lost over 6000 votes, while Riley's count didn't change. Siegelman's vote decreased by 1/3, roughly double the 18% vote flip. Gundlach concluded that GOP officials attempted to shift 18% of Siegelman's vote to Riley, but rather than subtracting that from Siegelman and adding it to Riley, they erroneously added it to both candidates. Overnight, the election officials fixed the vote flip and subtracted Siegelman's votes as intended. Siegelman's change from an 18% increase to an 18% decrease caused the 1/3 drop between the first and second returns.

If Siegelman's votes in Baldwin County weren't flipped, he would have gotten 15000 votes to Riley's 28000 votes. This would have been enough for Siegelman to win the governor's race.

Blocked recount

Democrats quickly moved to recount the paper ballots, but Republican attorney general William Pryor ordered the ballots sealed. Unless Siegelman challenged it in court, it was illegal for the counties to conduct a recount.[2] Since the attorney general would be in charge of a recount, Siegelman expected Pryor to manipulate it in Riley's favor, and opted to concede rather than be embarrassed by a sham recount.[4]

Political prosecution

Main article: Don Siegelman prosecution

Culprit identified?




  1. Washington Post, "Post-Election Alabama Is Seeing Double", 2002/11/08 - post-election confusion over the winner
  2. 2.0 2.1 Raw Story, "The permanent Republican majority: Part one: How a coterie of Republican heavyweights sent a governor to jail", 2007/11/26
  3. James H. Gundlach, "A Statistical Analysis of Possible Electronic Ballot Box Stuffing, The Case of Baldwin County Alabama Governor’s Race in 2002", 2003/04/11
  4. Raw Story, "The permanent Republican majority: Daughter of jailed governor sees White House hand in her father's fall", 2007/11/27
  5. Glynn Wilson, "How the 2002 Election Was Stolen in Bay Minette", 2007/06/20 - alleged role of Riley staffer Dan Gans in committing the fraud
  6. Alexander Strategy Group (ASG) bio of Dan Gans - says that he was behind a "ballot security program" critical to Riley's victory
  7. Gulf Coast Newspapers, "The changing of the guards: Bay Minette, election night" by Steve McConnell, 2007/07/20 - denial by Gans that he rigged the election or was even in Baldwin County on election night

External links

  • Fraud perpetrators
    • Jackson Free Press, "Secrets & Lies: GOP Accused of Political Prosecutions", 2008/06/04: "Kitty McCullough—also known as Kelly Kimbrough—was Rove's business partner at his political consulting firm K. Rove & Company in Alabama. The state Republican Party gave her credit for finding the irregularity that Siegelman describes as vote theft. The other operative, Dan Gans, is a self-described electronic ballot security expert who later went to work for a Tom Delay and Jack Abramoff-related company, The Alexander Strategy Group—which has been implicated in the Abramoff scandal—and had close ties to now Mississippi Gov. Haley Barbour. Gans claimed credit on his Web site for finding the votes that delivered the election to Riley. [...] Gans and Kimbrough were two people, among others, who lingered around a Baldwin County voting machine after the reporters had declared victory for Siegelman and gone home. The very next day, about 5,000 votes from that county magically switched to Riley. The odd thing about the new votes was that many of the voters choosing Republican Riley had chosen Democrats on down-ticket races."
    • Glynn Wilson, "Rove Issues Non-Denial Denial in Siegelman Case", 2009/08/12: "When asked about other e-mails showing he communicated with Kitty McClullough at the RNC and Sara Taylor about “doing some digging” on the Riley-Siegelman race, he sidestepped the question, but admitted the information he got came from Toby Roth, Riley’s chief of staff. [...] When asked again if he discussed Siegelman with Riley or Canary, Rove dodges the question. “I may have had conversations with Kitty McCullough,” he says."