A story in today’s Arizona Daily Star was brought to my attention today by some colleagues. The money quote is here:
Of the 1.7 million ballots cast, nearly 300,000 were early ballots that weren’t early after all, but folks dropping them off in unprecedented numbers at their polling places, which then triggered a new round of verification to make sure the ballots were authentic. There were another 84,000 provisional ballots that had some problems with registration or identification.
These strike me as really big numbers: 17% of ALL ballots in AZ were no-excuse ballots that dropped off at a precinct place on election day? More than half of Arizonans cast early ballots, so this translates into at least 35% of no-excuse ballots dropped off at a precinct place on election day, a number that is a low estimate since it divides the no-excuse ballots by ALL early votes, including early in person (the SoS website does not discriminate between the two modes).
These would be shockingly high figures, if they were right. In comparison, Los Angeles County reported to me that 14.9% of all no-excuse absentee ballots cast in 2008 were dropped off at a precinct place on election day, and those ballots constituted 3.6% of the total ballots cast. As I have reported in the past, 15%-30% of ballots in Oregon are hand-delivered (or arrive through the mail) at county offices or a library drop box on election day, but these are not local precinct polling places.
A bit of detective work has revealed that the story cannot possibly be accurate, unless there has been a massive population shift in Arizona and Maricopa no longer constitutes more than half the state. According to Maricopa, they sent out 866,440 no-excuse ballots and the return rate was 77% (approximately 667,000 returned). Of the 665,065, 117,129 were returned to a polling place on election day, or 17.6%, a figure similar to what Los Angeles County experiences. That leaves less than 400,000 additional early votes cast in the state, according to the SoS office. It just doesn’t seem realistic that more than half of those were no-excuse ballots returned on election day.
Why does this matter? It does for two reasons. First, election officials are losing their jobs over “slow counts.” As long as states allow citizens to drop their ballots at the polling place on election day, or accept a post mark rather than delivery by election day, slow counts are inevitable. In Arizona’s case, voter ID requirements leads to a high number of provisional ballots, only further slowing the count. In a report issued by the Pew Center on the States, Arizona ranked first in the nation in provisional ballots, with 6.61%, in 2008. I suppose reducing that number to 5% in 2010 might be seen as progress, although until the 2010 EAC Election Day Survey report is issued, we won’t have a good sense of how provisional ballot rates vary from presidential to midterm years. Still, voter ID inevitably produces higher numbers of provisional ballots. That may be an acceptable trade off, but don’t then complain about slow counts.
Second, we have noted in previous work that “early” voting is something of a misnomer, and most voters actually cast their ballots in the last week of the election, undercutting recent claims that early voting results in uninformed voters and has “stolen” election day. In Arizona and California, about 15% of “early” voters drop ballots off on election. In Oregon, a quarter to a third do so. And we have shown in other states, that well under half of the ballots are cast earlier than a week before election day.
So much for “stealing” election day.