Don’t Put Too Much Stock in Early Voting Numbers

  • submit to reddit

Voters arrive and depart during early voting at a polling place at the Wicomico County Youth and Civic Center in Salisbury, Md. October 2012. (AP Photo/Alex Brandon)

With candidates in the final push to win over swing voters in key states, it’s easy to forget that many Americans have already voted by absentee ballots or at an early voting location. In Ohio — the state that Nate Silver’s 538 blog says is about 50 percent likely to play the deciding role in the election — more than a million voters had cast their ballots by Oct. 26 (the latest date for which the Ohio Secretary of State has released data). In Florida, another key swing state, nearly 3.5 million votes have been cast. In total, over 24 million votes have already been cast nationwide.

Journalists, commentators and the campaigns are eagerly interpreting the data, but The Atlantic’s Molly Ball writes that it’s not a simple matter of seeing how many votes have been cast for each candidate so far.

In many states, election officials disclose how many Democrats and Republicans have voted thus far. We don’t know who they’re voting for, but in most states, this alignment is a good proxy for the candidates. (Then there’s the mystery of those voters who aren’t affiliated with a party. In polls, independent voters have generally favored Romney, leading his campaign to claim an edge in this category, but for the purposes of analyzing early voting it’s impossible to tell.) It’s important to consider which party has historically had the early vote advantage — Democrats, in most states — and whether early voting makes up a substantial amount of the vote, which varies from state to state.

Ball goes on to note that early voting numbers are not necessarily an indication of who the likely winner might be, but may provide some clues as to which campaign is doing a better job of getting out the vote. In North Carolina, for example, Democrats are leading the early vote by a significant margin — 48 percent of early votes come from Democrats, whereas only 32 percent come from Republicans. But Ball observes that Democrats might look upon those percentages as cause for concern.

Democrats’ early-vote lead in North Carolina certainly looks formidable. But in 2008, Democrats won the North Carolina early vote by an even wider margin, 51 percent to 30 percent, and only carried the state by 14,000 votes — less than half a percentage point. They can’t afford any drop in early voting.

George Mason University Professor Michael McDonald who maintains an excellent website tracking early voting data cautions that although early voters’ registered party affiliations are reported, their actual votes are not. And voters do not always vote for their party’s candidate.

Breakdowns of early voters by party registration, where available, are not votes for president. We do not know who a person registered with a party voted for; these early votes are tabulated on Election Day or after, depending on state law. That said, we might infer that a person registered with a party is likely to support their party’s presidential nominee.

Regardless, both campaigns are already saying that early voting numbers look encouraging for their candidate, and both Joe Biden and Paul Ryan have been encouraging voters to cast their ballots early.

Hurricane Sandy may also impact early voting. As many communities in Pennsylvania, Connecticut, New York and New Jersey struggle to recover and get their polling places open and operational, early voting by mail may be a more convenient option. In the wake of the storm, New York and New Jersey extended deadlines for voting by absentee ballot to today. The Boston Globe reports that the storm has also disrupted in-person voting in Maryland, West Virginia and North Carolina.

  • submit to reddit