American Prospect - A new report from the Voting Technology Project .... shows that votes cast by mail are significantly less likely to be counted than those cast in person. The report has serious implications given recent trends toward more and more mail-in ballots. Voting by mail has grown from less than 10 percent of ballots cast in 2000 to 17 percent in 2010. Two states, Oregon and Washington, conduct elections exclusively through the mail, while several others, including California and Colorado, allow voters to become permanent absentee voters, automatically getting a mail-in ballot every year.
In 2008, 800,000 mail-in ballots were rejected by election workers for one problem or another. Another 3.9 million were requested by voters but never received, while 2.9 million were sent to voters but never made it back to election officials. In total, as many 7.6 million votes, 21 percent of those requested, may have “leaked” out of the system before the votes were counted. It’s still the case that the total number of mail-in ballots cast and rejected is small—around 2 percent of those requested—but the gap in accuracy is certainly cause for concern. And in a tight election, those uncounted ballots could make a difference.
“It continues to surprise me,” says Charles Stewart, a political science professor at MIT and one of the authors of the report, ”that with all of the growth in voting by mail, that there has been surprisingly little curiosity about how accurate the voting mode is when you vote by mail.”