Partisan actors attempted to probe the accuracy of New Jersey’s voter rolls by comparing them with death records and with the rolls of other states. The reports led to inflated claims of widespread fraud, of the sort commonly used to support restrictive identification requirements for voters at the polls. We examined each of the allegations of fraud by individual voters—the only sort that ID could possibly address—to uncover the truth behind the assertions.
The allegations yielded only eight substantiated cases of individuals knowingly casting invalid votes that counted—eight voters who voted twice. Given the number of votes cast in these elections, this amounts to a rate of 0.0004%. None of these problems could have been resolved by requiring photo ID at the polls.
The analysis below examines the allegations of fraud in more detail.
The rate of substantiated fraud:
- The allegations of fraud related to the 2004 general elections, in which 3,611,691 votes were cast in New Jersey.
There were eight substantiated cases of individuals knowingly casting invalid votes that counted—eight voters voting twice. This amounts to a rate of 0.0004%. None of these problems could have been resolved by requiring photo ID at the polls.
- Even given allegations that were unsubstantiated, the rate of possible fraud remains low. The analysis below lays out the allegations, reasons to question each, and the facts that we know. But assuming that all 13,419 of the remaining cases in fact involved fraud—which is highly unlikely, given the methodological errors revealed in the study of double-voting—that would amount to a rate of 0.61%.
- 4,755 votes were alleged to have been cast in the names of dead voters in 2004, based on an attempt to match the first and last name and date of birth from voting records to death records.
- 4,397 individuals voted twice in New Jersey, and 6,572 individuals voted both in New Jersey and in either New York, Pennsylvania, Florida, North Carolina, or South Carolina, based on an attempt to match the first and last name and date of birth from one set of voting records to another.
Questioning the allegations:
- Dead voters: Attempts to match records of deaths to current voter rolls are often used to buttress claims that the dead are voting, but the matching enterprise is unreliable. Two list entries under the same name – even the same name and birthdate – may indicate different individuals. Even more often, the lists themselves are flawed: because of a clerical error or faulty information, an individual is marked as voting when she did not in fact cast a ballot, or an individual is shown as deceased when she is actually very much alive. Finally, it is possible that the match is accurate but reveals nothing illegal about the vote: the voter has died, yes, but after casting her ballot.
- Double voters: The fact that the same name is listed as voting on two different voting rolls does not necessarily mean that one person voted twice: often, what looks like the same person listed twice turns out to be two different individuals with the same name. In other cases, as with dead voters, clerical errors are responsible for the double listing.
- Dead voters: No follow-up investigation appears to have been published on the number of votes actually cast in the names of dead voters in 2004, if any. None of the allegedly dead voters actually voted in 2005.
- Double voters: Analysis of the list of alleged double-voters within New Jersey showed that 2,305 of the entries had different middle names or suffixes, or an error in the date of birth. Data errors in Middlesex county, and the statistical likelihood of finding two different individuals with the same name and birthdate, call into question much of the remainder of the list. Through examination of 60 original signatures on poll book materials, the New Jersey Republican State Committee claimed that eight voters were definitively identified as having voted twice.
Coverage by existing law: The federal Help America Vote Act (HAVA) requires states to create statewide electronic voter registration lists, and to coordinate those computerized lists with agency records on death in order to remove ineligible voters. Although the obligation to remove deceased voters from the rolls predated HAVA, the computerized registration rolls—if implemented with suitable controls for accuracy—offer a new and efficient means to do so statewide. Like most states, New Jersey did not have a HAVA-ready statewide database up and running in 2004, but once it does, the database should allow the state both to eliminate duplicate registrations within the state and to cut down on the number of deceased citizens who are still on the rolls.
Cynthia Burton, No Beyond-the-Grave Balloting Cited, Phila. Inquirer, Nov. 9, 2005.
James A. Quirk, Local County Boards Question GOP Voter-Fraud Study, Asbury Park Press, Sept. 17, 2005.
David W. Chen, Among Voters in New Jersey, G.O.P. Sees Dead People, N.Y. Times, Sept. 16, 2005.
Joe Donohue, GOP Claims Voter Fraud Widespread, Star-Ledger, Sept. 16, 2005.
Bonnie Pfister, Review Election Rolls, GOP Urges, Phila. Inquirer, Sept. 16, 2005.
Tom Hester, Jr., GOP: Voter Rolls Are Rife with Fraudulent Names, The Times (Trenton, N.J.), Sept. 16, 2005.
Mitchel Maddux, GOP Calls for Probe of State’s Election Rolls, The Record (Bergen County, N.J.), Sept. 16, 2005.
Letter from Mark D. Sheridan to Peter C. Harvey, Attorney General, Sept. 15, 2005.
Brennan Center for Justice & Dr. Michael McDonald, Analysis of the September 15, 2005 Voter Fraud Report Submitted to the New Jersey Attorney General, Dec. 2005.
New Jersey Office of the Attorney General, Division of Elections, 2004 General Election Results.