What Constitutes an Election Audit


Guest Author – Richard Hayes Phillips, Ph.D.: Ohio citizens conducted, under my direction, a genuine audit of the 2004 presidential election. This was no mere “spot check” of randomly selected precincts, and no mere “recount” of the same ballots previously run through the electronic tabulators.

We learned to ask for everything: ballots, poll books, voter signature books, ballot accounting charts, packing slips, and invoices. We asked to see all the ballots, whether voted, spoiled, or unused. And we always asked to photograph the records, so that I could analyze them with painstaking accuracy, and reexamine the same records when necessary.

We lacked the element of surprise, as a list of requested precincts was almost always demanded in writing and well in advance. But I picked the counties, and I picked the precincts, and I rarely said how or why. “What are you looking for?” I was often asked. “I don’t know,” I would reply. “This is an audit.”

When the IRS audits your tax returns, you don’t get to pick the year, or decide which records to show them. They want to see everything. And so did we.

When election results have been altered, this will almost always be apparent at the precinct level. Either the numbers will be at variance with long-established voting patterns, or inexplicable combinations of choices will be attributed to the same voters on the same day, or both. Voter turnout, that is, the percentage of registered voters casting ballots, may be suspect, either too high or too low. The percentage of ballots recorded as having no choice for the office, equal to undervotes plus overvotes, may be anomalously high or low. Based upon these criteria, we audited the most suspect precincts.

All of the records we requested are important. It is rightly the responsibility of election officials to verify the accuracy of the elections they administer.

The ballot accounting charts for each precinct should state the number of ballots received at the start of the day, which should match the number on the itemized packing slip from the printer who supplied the county with all its ballots. That same chart should state the total number of “voted” ballots, which should equal the number of names in the poll book and the voter signature book. It should state the number of “spoiled” ballots, which should match the number of altered ballot stub numbers recorded in the voter signature book. And it should state the number of “unused” ballots remaining at the end of the day, which, when added to the number of “voted” and “spoiled” ballots, should equal the total number of ballots received at the start of the day. Without these records there is no way to tell if the ballot box contains too many ballots, or too few.

Ballot stubs are numbered strips of paper attached to each ballot. The stub number for each ballot issued, both “voted” and “spoiled,” should be recorded by a poll worker right next to the voter’s name in both the poll book, written by the poll worker, and the voter signature book, signed by the voter. The ballot stub should be torn off and placed into the ballot box separately, to protect voter privacy and the right to a secret ballot. The numbers on the torn-off stubs should match the stub numbers in the poll book and the voter signature book, and the numbers on the stubs still attached to the unused ballots should not; and all the stubs, and all the ballots, whether voted, spoiled, or unused, should be preserved. Without these records there is no way to tell if the ballots run through the electronic tabulator are the same ballots issued to the voters.

In Ohio, Boards of Elections are at liberty to “remake” ballots at their discretion, ostensibly so that the voter’s intent will be accurately recorded by the electronic tabulator. In the counties we audited, the number of “remakes” or “duplicates” ranged from a mere handful to more than one percent of the total ballots cast in the entire county. The original “spoiled” ballots which the “remakes” allegedly duplicate are supposed to be preserved. We never saw any of them. Without these records, there is no way to tell if the “remakes” are legitimate.

The subsets of regular, absentee, and provisional ballots in each precinct are also supposed to match the corresponding numbers of names recorded in the poll book and the voter signature book. If the books do not indicate which absentee ballots were returned by the voters and which were not, and which provisional ballots were approved and which were not, another opportunity arises for alteration of the vote count.

The ballots for each precinct must be kept in the same sequence in which the auditor found them. Failure to do so can compromise the evidence. Long consecutive runs of ballots for one candidate or another are proof of hand sorting, for which there might be no legitimate reason. Abrupt changes in voting patterns partway through the stack of ballots may be indicative of ballot tampering, especially if there is a marked increase or decrease in “ticket splitting.” This is why “whole ballot analysis” is essential. The combinations of choices attributed to individual voters on each ballot must be examined, not merely the contest being investigated.

Ballots from numerous counties must be examined. Unless this is done, there is no frame of reference, and there is no way to tell if ballots are counterfeit. Likewise, all the marks on the ballot must be examined, to see if one or more of the marks are made by a different hand than the others. Such forgeries can be a method for spoiling the ballot by turning the voter’s choice into an “overvote,” or by turning an “undervote” into a vote for the candidate desired by the election riggers.

One cannot overstate the importance of the chain of custody for the ballots, as it is here that the opportunity for election rigging arises. Lapses in the chain of custody after the ballots leave the polling place on Election Night provide the opportunity for ballot tampering prior to tabulation, in which case a subsequent hand count will nicely match the tabulator count. Lapses in the chain of custody after the ballots are tabulated provide the opportunity for ballot substitution in order to get the ballots to match a rigged tabulator count. And the greater the number of “extra” ballots ordered by the Board of Elections, above and beyond what could possibly be needed to accommodate all the voters, the greater the margin by which the vote count can be altered. All that is needed to cover the tracks is to destroy the unwanted ballots and the unused ballots, or to leave the “extra” ballots off the invoice and the packing slip in the first place.

Despite the numerous methods of ballot tampering practiced in Ohio, doing away with paper ballots is not the solution. Quite the contrary; the fact that eighty-five percent of the votes in Ohio in the 2004 election were cast on paper is what made the fraud detectable in the first place, whereas electronic voting with no paper record makes election fraud undetectable. What made ballot alteration and ballot substitution possible in Ohio were the breaks in the chain of custody; and what allowed the 2004 election to withstand the initial court challenge was the fact that investigators were not allowed to examine the ballots until 2006.

The preconditions for any crime are motive, means, and opportunity. In case of election fraud, the motive will always be provided by the desire to win the official count, and the means will always be provided by whatever voting method is used. The only way to prevent election fraud is to prevent the opportunity.

In my judgment, based upon three years’ experience auditing a rigged presidential election, the solution is this: paper ballots, counted by hand, in full public view, at the polling place, on Election Night, no matter how long it takes. In this way the counting takes place before any chain of custody questions have arisen, which effectively prevents the opportunity for wholesale election fraud associated with central tabulation. If this seems old-fashioned, so be it. When one is on the wrong path, a step backward is a step in the right direction.

Make them steal elections the old-fashioned way, by altering ballots, destroying ballots, or stuffing the ballot boxes right at the polling places, in precinct after precinct. This requires the collusion of large numbers of poll workers, both Republican and Democrat, and runs the risk of exposure at any polling place where we, the people, are watching.

Presented at the Building a New World Conference – Radford, Virginia, May 24, 2008

Copies of “Witness to a Crime” can be obtained from the author: richardhayesphillips at gmail dot com


Richard Hayes Phillips holds a B.A. in Politics and Geography from the State University of New York at Potsdam, M.A. degrees in History and Geography from the University of Oklahoma, and a Ph.D. in Geomorphology from the University of Oregon.  He has taught twelve different courses at seven colleges and universities. He is the author of many books and professional papers, including “Witness to a Crime: A Citizens’ Audit of an American Election” (Canterbury Press, Rome, New York, 2008), and most recently a historical and genealogical trilogy on white child slavery in colonial Maryland and Virginia: “Without Indentures,” “Birth and Shipping Records,” and “The Search for Survivors.”  These are standard references in more than one hundred libraries and archives in the United States, Canada, England, Scotland, Ireland, Australia, and New Zealand.



Share This