4:46 p.m. | Updated At just before 2 a.m. on Wednesday, Matt Strawn, the Iowa Republican Party chairman, addressed reporters who had spent the night following the historically close Republican caucus race between Mitt Romney and Rick Santorum. The margin separating the candidates was small — eight votes — but he was prepared to declare a victor. “Congratulations to Governor Mitt Romney, the winner of the 2012 Iowa caucuses,” Mr. Strawn said.
Mr. Romney’s victory is unofficial — the counties have up to two weeks from the caucuses to send their final certified results to the state party. However, there is no provision for a recount in the caucuses, and the campaign that might have the most interest in pursuing one — Mr. Santorum’s — is making no effort to challenge the results.
Still, given Mr. Romney’s exceptionally small margin of victory, a single discrepancy could potentially reverse the outcome. On Wednesday, a voter in the town of Moulton in Appanoose County claimed to have found one.
The voter, Edward True, signed an affidavit stating that he had helped count the vote after the caucus at the Garrett Memorial Library in Moulton. Mr. True claims that the results listed on the Google spreadsheet maintained by the Iowa Republican Party differed substantially from the count that had been taken at the caucus site. Mr. Romney had received only 2 votes in his precinct, Mr. True’s affidavit said, but had been given credit for 22 by the state. That would be enough to flip Mr. Romney’s 8-vote victory into a 12-vote win for Mr. Santorum.
In an interview Friday, Mr. True said that he is a Ron Paul supporter.
Mr. Strawn issued a statement late on Thursday night on behalf of the Iowa G.O.P., which said that the party would not respond in detail to reports of irregularities, but that it had no reason to believe that Mr. True’s allegation would change the outcome.
One reason for Mr. Strawn’s confidence may be that there is a verification process set up during the caucuses. Each candidate has the option of assigning a representative to each precinct who must sign off on the results before they are reported to the state.
Whether Mr. Santorum’s campaign had such a representative at Mr. True’s precinct, designated by the state as Washington Wells, is unclear, but the better-organized campaigns typically have representatives at most of the state’s 1,774 precincts.
I spoke with Mr. Santorum late on Thursday evening when I encountered him in the lobby of the Manchester, N.H., hotel where he is staying. Mr. Santorum was in a jovial mood, joking about his sweater vests and the lack of sleep he had been getting. “I stayed up until CNN was on the morning show,” he recalled of the caucus night.
Mr. Santorum expressed a cheery indifference as to whether he had technically won the caucuses. “I always said in Iowa that I had to beat Perry and Bachmann,” he said. “I didn’t really ever think I was going to win, given where I was two weeks out. It just was fun watching it going up and down.”
Mr. Santorum said that his campaign had not put any significant effort into looking at the precinct-by-precinct results in detail, nor was it interested in doing so.
Jesse Benton, Mr. Paul’s national campaign chairman, said on Friday that he did not know anything about Mr. True’s allegation. But he noted that Mr. Paul’s campaign had ample representation at both the state and precinct levels during the vote-counting process. “We’ve always been very satisfied that they are playing above-board,” Mr. Benton said.
Earlier on Thursday, Mr. Santorum conducted an interview with Greta Van Susteren of Fox News during which he noted that discrepancies had run both ways during the Iowa caucuses. In the interview, Mr. Santorum said that although there may have been a 20-vote mistake in Mr. Romney’s favor in one precinct, there was a 21-vote mistake in his favor in another precinct, essentially canceling it out.
When I asked him about the 21-vote discrepancy, Mr. Santorum told me that he was not aware of which specific precinct it had occurred in. Instead, he said, he was recalling information from prior conversations he had with Iowa Republican officials.
In fact, there were several discrepancies during the vote-counting on Tuesday night, some of which had favored Mr. Santorum. However, these discrepancies were resolved and corrected before Mr. Strawn declared Mr. Romney the winner.
I had been downloading versions of the state’s vote tally at various points on Tuesday evening and early Wednesday morning. The first version I downloaded reflected the state’s count as of 11:14 p.m. Iowa time on Tuesday, at which point results from 1,723 of 1,774 precincts had been tabulated.
In 10 of these 1,723 precincts, the results in the spreadsheet changed between 11:14 p.m. and 2 a.m., reflecting additional verification and validation efforts. These changes added a net of 11 votes to Mr. Romney’s total while subtracting a net of 33 votes from Mr. Santorum’s count. If not for these changes, Mr. Santorum would still have led the vote count on Wednesday morning.
However, it is not clear that there are additional known discrepancies in Mr. Santorum’s favor, above and beyond the ones that have already been detected and fixed. Thus, the potential discrepancy described by Mr. True could in fact be enough to reverse the result.
It is possible to apply some statistical scrutiny to Mr. True’s claim. How plausible is it, for instance, that Mr. Romney would have received just two votes out of 53 in the precinct, as Mr. True reported?
In fact, Mr. Romney performed very weakly in the other precincts in Appanoose County, which is rural and conservative. In the Centerville W1 precinct, for instance, Mr. Romney received just two of 39 votes. He got just four of 43 votes in the Caldwell precinct, and none of the 40 voters in the Union precinct cast a ballot for him.
Another unusual facet of the vote count in the Washington Wells precinct is that the state recorded six votes for an obscure candidate, Buddy Roemer, even though Mr. True’s affidavit claimed that there had been no votes for him there. (Mr. Roemer received just 31 votes statewide.) Over all, Mr. True’s statement said, 53 caucusgoers had cast votes in his precinct, rather than 79 as reported by the state. The 26-vote difference reflects what he says were 20 excess votes for Mr. Romney and six for Mr. Roemer.
Is there any way to tell whether either 53 votes or 79 votes is an inherently more believable turnout in the Washington Wells precinct?
Neither voter registration statistics nor the results of the 2008 caucuses are available on a precinct-by-precinct basis. However, precinct-level results of the 2008 general election are provided on the Iowa secretary of state’s Web site, and the number of votes for John McCain can be used as a proxy for the number of Republicans.
In the 2008 general election, 212 voters cast ballots for Mr. McCain in the Washington Wells precinct. A 79-voter Republican caucus turnout in Washington Wells would represent 37 percent of this total. Statewide, by comparison, Republican caucus turnout was equal to 18 percent of Mr. McCain’s vote total.
The 53-voter turnout as instead claimed by Mr. True would represent 25 percent of Mr. McCain’s vote total. That would be somewhat more in line with statewide averages, as well as the 23 percent turnout reported elsewhere in Appanoose County. Still, caucus turnout can be highly variable from precinct to precinct, so this evidence is far from definitive.
There is also another, more striking oddity in the turnout estimates in Appanoose County. One precinct there, Pleasant Franklin, is still being reported by the Iowa Republican Party as having had no turnout at all for the caucus. This is despite the fact that Pleasant Franklin was listed as a valid caucus site on the Republican Party’s Web site, and that 132 voters cast ballots for John McCain there in the general election of 2008.
In addition to Pleasant Franklin, there are seven other precincts throughout Iowa that are still being reported as having had no Republican turnout. These precincts are listed in the state party’s spreadsheet, but contain no votes.
It is important to note that apparent oddities and anomalies are common whenever votes are counted. Some prove to be benign, while others are genuine mistakes, but they rarely make much difference either way.
Eight votes, however, is not a lot of ground for Mr. Santorum to make up, so the outcome could easily change before the vote is certified. For now, the caucuses are probably best thought of as still being too close to call.
This post has been updated the reflect the fact that Mr. True is a supporter of Mr. Paul.