Take the 2-minute tour ×
Skeptics Stack Exchange is a question and answer site for scientific skepticism. It's 100% free, no registration required.

The issue of voter fraud comes up very often at election times. In the media, it is nearly always pictured as a really important and urgent problem. It ranges from whether photo identification should be required, to rigged voting machines, to corrupt officials, to blatant lies (such as Was there 100% turn out and 100% Obama support in Cuyahoga County in 2012 Presidential Election?)

However Stephen Colbert stated on his show that voter fraud is for all intents and purposes non-existent in the United States, citing a MinnPost article which states:

“There is absolutely no evidence that [voter impersonation fraud] has affected the outcome of any election in the United States, at least any recent election in the United States,” Schultz said.

So, is the problem way overblown? Is there even a problem?

share|improve this question
outside of Chicago its probably true. –  Ryathal Nov 13 '12 at 13:37
I have seen several cases of it in local elections. But the scale required to impact a federal election makes it impractical and usually irrelevant. But the things that affect voters the most tend to be the result of local elections rather than state and national elections. –  Chad Nov 13 '12 at 16:04
@Chad why would it be impractical? –  Tass Nov 13 '12 at 17:56
Non-existent ≠ insignificant. It certainly exists. It probably isn’t significant. –  Konrad Rudolph Nov 15 '12 at 15:02
Where did Colbert make that claim? Could you provide a link? –  Christian Nov 16 '12 at 0:32

2 Answers 2

up vote 21 down vote accepted

Well, after a quick search it seems that Stephen Colbert is probably correct.

First, it's good to mention that according to the Department of Justice most Voter Fraud will fall under state jurisdiction and not Federal jurisdiction unless there are threats or discrimination based on race, religion, or national origin.

There is a popular figure oft-cited by The Daily Show and other sources who make arguments against certain Voter ID laws, which is 10 cases of "in-person" Voter ID fraud over the last decade or so. This seems technically correct, according to this article by ABC:

Over the past decade Texas has convicted 51 people of voter fraud, according the state's Attorney General Greg Abbott. Only four of those cases were for voter impersonation, the only type of voter fraud that voter ID laws prevent.

No doubt you noticed, though, that Stewart and other critics of Voter ID laws specifically mention cases that Voter ID laws would have prevented - although the total amount of Voter Fraud is higher.

Is it enough to matter? Well according to Google's statistics the difference in votes between the candidates was 3,378,662. A much closer race is the infamous 2000 Elections, where the FEC's statistics show a difference of 350,428 between the Democrat and Republican candidates.

It would have taken 7,009 cases of voter fraud per state in 2000 to make that gap, and over 67,500 cases of voter fraud per state to close the gap in 2012. I could not find a source for the total cases of voter fraud, but one of PolitiFact's articles attempts to address this:

To get the data, News21 reporters sent records requests to elections officers in all 50 states seeking every case of fraudulent elections activity, including registration fraud, absentee ballot fraud, voter impersonation fraud and casting an ineligible vote. News21 said it received no useful responses from several states. With some states, including Massachusetts, Oklahoma, South Carolina and South Dakota, the cases included in the database came from a survey of alleged election fraud conducted by the Republican National Lawyers Association. And in some states, some but not all local jurisdictions responded, and some responses were missing important details about each case. Despite those issues, News21 defends its work as "substantially complete" as the largest collection of election fraud cases gathered by anyone in the country.

They also go on to say that News21 found 307 cases of Voter Fraud (en totale, including Impersonation) in Georgia - which is 6x higher than Texas and over 10x higher than Pennsylvania (who News21 said had 29 cases). The Republican National Lawyer's Association challenged the analysis, saying Georgia had 375 cases.

Given that I've no baseline or average to work from, but Georgia appears to be one of the more rampant offenders - every state in the nation would have to have eighteen times the amount of voter fraud that Georgia had in a single year to make up the gap in the 2000 elections. It would take over 9,000-fold Georgia's cases per state in 2012 alone to have made up the gap in the 2012 elections.

Basically, even if all states had as much voter fraud as the RNLA said Georgia had per year (which they don't), it would still only represent .0155% of the total vote.

I'd agree with Stephen Colbert and other pundits; for all intents and purposes, Voter Fraud is practically a non-issue.

share|improve this answer
Thanks. I also wanted to add that in your answer, you assume that 100% of the voting fraud is committed by the losing side, which is certainly possible but not likely. It is more likely the other side commits it too, thereby evening it out a little bit and dramatically increasing the instances of voter fraud needed to change the results of the election –  Andreas Bonini Nov 13 '12 at 15:36
@Chaud: I think MCM understands your argument; he's simply asking you to prove it. He did prove his. If we just go around and say unsupported statements framed as absolute truth, like you are doing, we are really no better than the press and politicians who made voter fraud a huge deal for the 2012 presidential election. This is the opposite of what this site is about. –  Andreas Bonini Nov 13 '12 at 17:36
the one major problem I see with this answer is that its basing everything on the number of cases reported, which is the lower bound for actual amounts of fraud. Also it seems this answer equates one case of voter fraud to one vote, is that actually how it works, or is one person submitting 100 fraudulent votes only a single case and not 100 cases? –  Ryathal Nov 13 '12 at 18:22
@MCM I don't think I know the extend of US voting fraud. As a result I'm not in a position to answer the question. Your answer pretends to know the extend of US voting fraud. It's assumes without any argument that the US is good at detecting voting fraud. It also overestimates the amount of votes that you need to change an election by ignoring the fact that someone who wants to manipulate would focus on swing states. It further suggest that the threat model is wrong. –  Christian Nov 16 '12 at 16:05
Please keep the comments on topic - That is, about this answer and how it can be improved. Most definitely avoid discussing the question and other possible alternatives here. –  Sklivvz Nov 16 '12 at 16:21

Hans A. Von Spakovsky (Senior legal fellow at the Heritage foundation wrote and article called Voter Fraud Is a Proven Election Manipulation Tactic

The Supreme Court answered this question in 2008 when it upheld Indiana's voter ID law. "Flagrant examples of such fraud … have been documented throughout this Nation's history by respected historians and journalists," the court said, "[and] not only is the risk of voter fraud real but that it could affect the outcome of a close election.

But ask voters in Troy, N.Y., Lincoln County, W.Va., and Florida whether voter fraud is a real problem.Four local officials and party activists were convicted in 2011 of voter fraud in Troy for forging enough absentee ballots to "likely have tipped the city council and county elections" in 2009.

One of them who pled guilty, Anthony DeFiglio, told police that such fraud was a "normal political tactic."

In March 2012, the county sheriff and clerk in Lincoln County, W.Va., pled guilty to voter fraud... [T]he Lincoln County auditor was also found guilty of voter fraud in 2005.

Here are some other links:

share|improve this answer
Honestly, I'm not sure that this answer property address the question. Everyone pretty much agrees that voter fraud does happen; however, given the size of the population there isn't as much as you might expect and as far as I know there haven't been any reports of voter fraud influencing the results of an election. Can you find anything that address either of those? –  rjzii Feb 8 '13 at 18:06
voter fraud is for all intents and purposes non-existent - This shows that to be false. and it answers the question Is there even a problem? –  Chad Feb 8 '13 at 18:21
However, you also really haven't established on that on a large scale that it is a problem. For example, say we concede that it is a problem in elections where votes are measured in the 100's - not uncommon for small local elections - but what about cases were they are measured in the 1,000,000's? In other-words, in the United States you really only hear about voter fraud being a problem during national elections. When examined from that level, which is the level Colbert was referring to, is there still a problem? –  rjzii Feb 8 '13 at 19:09
I looked up the show and the article that was cited by Colbert in the shower, and updated the question accordingly. There might actually be two aspects to the question though. Colbert was referring specifically to voter-id fraud as opposed to general election fraud. In otherwords, ballot stuffing is a bigger issue than someone showing up and claiming to be someone else. –  rjzii Feb 8 '13 at 19:28
I followed the "Here are some other links" and none of them, with the possible exception of the last two (which do not have sufficient specifics), seem to involve more than a handful of fraudulent votes. I do not think those links support the contention that the fraud would have been statistically significant. –  Larry OBrien Feb 8 '13 at 21:28

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.