As more voters take advantage of the opportunity to cast ballots before Election Day, it seems reasonable to expect that turnout will grow as well, right?

Not necessarily, according to those who've studied it. The evidence of early voting's effect on turnout is actually quite mixed. Early voting can energize turnout, but it depends on the kind of election, the form of early voting available (vote-by-mail, "no-excuse" absentee, early-in-person, etc.) and a state's registration laws.

A team of University of Wisconsin political scientists analyzed 2008 election returns in each of the nation's 3,100 counties, controlling for variations in demographics, geography and political dynamics (ie. whether a county was in a battleground state). Turnout in counties with early voting was about three percentage points lower than in counties without it.

"Essentially what we concluded is counterintuitive, but early voting has a pretty consistent negative effect on turnout," said Kenneth Mayer, political science professor at Wisconsin. The findings, published in 2009, were supported by census data on the characteristics of voters and non-voters. They suggested that early voting attracts those most likely to vote regardless of whether an early option is available.

It seems implausible, given the record turnout (130 million) in 2008, which included a big spike in early voting, especially early in-person voting by African Americans. 

But Mayer and his colleagues identified two major factors that could keep early voting from having more of an effect on turnout. One is that voting in most states is still a two-step process: registration on one day, then casting a ballot, sometimes weeks or months later. Having the opportunity to vote early doesn't matter much if you're not registered.

Wisconsin researchers also cited previous studies suggesting that early balloting erodes the intensity of traditional get-out-the vote efforts in the final days, weakening "the stimulating effect it would otherwise have on non-voters."

There is one major exception, the Wisconsin scholars found. States that allowed voters to register and cast an early ballot on the same day in 2008 showed turnout levels higher than the national rate of 58 percent. A small number of states, including North Carolina, Iowa and Wyoming, have some form of early same-day and Election Day registration.

"If the motivation for election reform is increasing turnout, states should not look to early voting, especially on its own," their study concluded.

Paul Gronke of Reed College, one of the leading early vote scholars, said he's "not a big fan of the Wisconsin study," because it was limited to one year's worth of data. His own study a few years ago found that a combination of early in-person voting and "no-excuse" absentee balloting had a small but statistically discernible impact (about 2.6 percent) on turnout, but only during mid-term elections. He also said that early voting effects are most dramatic in local, off-cycle elections where vote-by-mail is employed. Gronke theorized that the absence of saturation press coverage and wall-to-wall television advertising that accompanies state and national contests, actually concentrates the attention of voters who receive mail ballots.