All of these restrictive measures take their justification from a perceived need to prevent "voter fraud." But there is overwhelming scholarly and legal consensus that voter fraud is vanishingly rare, and in fact non-existent at the levels imagined by voter ID proponents. That hasn't stopped many Republican lawmakers from crying "fraud" every time they're faced with an unfavorable election outcome (see also: McDaniel, Chris).
For reference, a round-up of the latest research is below. Let me know in the comments if I missed anything.
The Truth About Voter Fraud, by Justin Levitt of Loyola Law School. Levitt performed a wide-ranging analysis of alleged incidents of voter fraud across the U.S. "Usually, only a tiny portion of the claimed illegality is substantiated — and most of the remainder is either nothing more than speculation or has been conclusively debunked."
The Politics of Voter Fraud, by Lorraine Minnite of Columbia University. Minnite concludes that voter fraud is exceedingly rare, and that the few allegations in the record usually turn out to be something other than voter fraud: "a review of news stories over a recent two year period found that reports of voter fraud were most often limited to local races and individual acts and fell into three categories: unsubstantiated or false claims by the loser of a close race, mischief and administrative or voter error."
Fraudulent Votes, Voter Identification and the 2012 US General Election, by John Ahlquist and Kenneth R. Mayer of the University of Wisconsin, and Simon Jackman of Stanford. The authors conducted a survey experiment "to measure the prevalence of two specific types of voter fraud: repeat/fraudulent ballot casting and vote buying." Their conclusion: "The notion that voter impersonation is a widespread behavior is totally contradicted by these data."
Voter Identifications Laws, by Minnite again. "In 95 percent of so-called 'cemetery voting' alleged in the 2010 midterm election in South Carolina, human error accounts for nearly all of what the state's highest law enforcement official had informed the U.S. Department of Justice was fraud."
In 2011 a Wisconsin task force found sufficient evidence to charge 20 people with fraudulent voting in the 2008 elections. Most of these were felons who were ineligible to vote.
Kansas' secretary of state examined 84 million votes cast in 22 states to look for duplicate registrants. In the end 14 cases were referred for prosecution, representing 0.00000017 percent of the votes cast.
A 10-year 'death audit' in North Carolina turned up a grand total of 50 instances in which a vote may have been attributed to a deceased person, most likely due to errors made by precinct workers.
The New York Times examined five years of Justice Department records and turned up only 26 convictions of fraud by individual voters. "Many of those charged by the Justice Department appear to have mistakenly filled out registration forms or misunderstood eligibility rules." And many of the cases were "linked to small vote-buying schemes in which candidates generally in sheriff’s or judge’s races paid voters for their support." Overall, the Times concluded that "the Justice Department has turned up virtually no evidence of any organized effort to skew federal elections."
News 21, a student journalism project based at Arizona State University, analyzed 2,068 alleged fraud cases from 2000 to 2012 and found a total of 10 cases of alleged voter impersonation. They found that "while fraud has occurred, the rate is infinitesimal, and in-person voter impersonation on Election Day, which prompted 37 state legislatures to enact or consider tough voter ID laws, is virtually non-existent."
In striking down Wisconsin's voter ID law this spring, district judge Lynn Adelman wrote "The evidence at trial established that virtually no voter impersonation occurs in Wisconsin. The defendants could not point to a single instance of known voter impersonation occurring in Wisconsin at any time in the recent past."