elections

In the modern era of presidential campaigns, even with candidates’ every last word and action carefully scripted, there can still be surprises. This year, that surprise came in the form of an interloper named Sandy, a blustery blowhard whose campaign stormed up the East Coast and made bombastic statements all along the way, treating red and blue states with equal disrespect. In the process, Sandy emphatically impacted the race for the presidency. First, she forced both President Obama and Mitt Romney to alter their campaigns while she commanded the attention of most of the Atlantic seaboard. They both had to step away from key battleground states, as the president canceled campaign stops in Florida and Wisconsin and Romney called off events in New Hampshire, Iowa, Ohio, and Wisconsin. The negative rhetoric between the two candidates was toned down while both focused their attention on the storm and its victims.

Sandy also caused the cancellation (and subsequent extension) of early-voting days and interrupted training of polling officials in some states, and her impact on voting day may be considerable even though her whirlwind tour is over. One potential effect could be that fewer voters will be able to get to the polls in hard-hit portions of the Northeast and mid-Atlantic, either due to transportation issues, damage to polling venues, power outages, or other hardships. This probably won’t change the election’s outcome, as Obama has a safe lead in most of the affected states. However, his popular vote might be adversely impacted, and thus  accentuating the significance of the electoral college. Additionally, polling and media coverage of the campaign has been cut back. Some organizations have temporarily suspended their polling, and with the news media tabbing Sandy as the new “it girl,” the presidential candidates were sent to the fourth estate’s back burner, at least for a few days, as notedby media reporter Howard Kurtz:

Still, the greatest impact may be on public attention. Go to any news website or flip on any cable news channel, and Sandy is the dominant story by far. No one is talking about tax cuts or unemployment or immigration. Television has a tendency to overhype major storms, but given the breadth and destructive power of Sandy, the saturation coverage seems to match its magnitude.

It’s also worth considering what effect Sandy will have on voters’ thoughts about climate and where the two candidates stand on climate change. New York City Mayor Michael Bloomberg cited Sandy in particular and climate change in general in his endorsement of Obama:

The devastation that Hurricane Sandy brought to New York City and much of the Northeast—in lost lives, lost homes and lost business—brought the stakes of Tuesday’s presidential election into sharp relief. . . .

Our climate is changing. And while the increase in extreme weather we have experienced in New York City and around the world may or may not be the result of it, the risk that it might be—given this week’s devastation—should compel all elected leaders to take immediate action. . . .

When I step into the voting booth, I think about the world I want to leave my two daughters, and the values that are required to guide us there. The two parties’ nominees for president offer different visions of where they want to lead America. . . .

One sees climate change as an urgent problem that threatens our planet; one does not. I want our president to place scientific evidence and risk management above electoral politics.

If many other voters share Bloomberg’s perspective that immediate action on climate change is necessary and the President is more likely to take that action, then Barack and Mitt may come to see Sandy as the October surprise that decided the election.

Predicting–as opposed to actually voting in–elections has become a national past-time, if one is to judge the media’s obsession with who’s going to win what in today’s midterm contests in the United States. And what better way to make predictions than to ponder the weather map?
Tony Wood of the Philadelphia Inquirer, however, disputes the oft-cited connection between weather and election results. In a post last week on “Weather, Democracy, and Mythology“, he looked into the old theory that rain dampens voter turnout and concluded that it doesn’t hold water:

Consider the 2004 presidential election. Recall that it all came down to Ohio in a close race between President Bush and Democratic nominee John Kerry.
That Election Day was a nasty one all over the state. It rained almost everywhere. The result? The voter turnout in Ohio was believed to be highest in at least 40 years. Some folks were said to be waiting up to nine hours to vote.
We did our own analysis of 30 years of election returns and weather in Philadelphia found no evidence of a link between the two. It rained on half of the Election Days with the 10 highest turnouts, while seven of the 10 lowest-turnout days were rain-free.

What about the oft-told story of the 1960 election? In his classic, The Weather Factor, meteorological historian David M. Ludlum claimed that rain in Illinois (on an otherwise mostly fair day across the country) hindered Nixonian rural voters more than Kennedy liberals in Chicago? Wood counters with political analyst Terry Madonna of Franklin & Marshall University, who says weather took a back seat to the behind-the-scenes intervention of Mayor Richard Daley:

“It [the rain] didn’t matter,” said Madonna, “because Daley had those votes already counted.”

We’re not sure how that explanation is logical (given that Ludlum’s theory is more about the lack of rural voters than about any surge in urban voters), but more specifically it would be great to see more of Wood’s 30-year study. In the meantime, the one recent study, by political scientist Brad Gomez and colleagues, quantifies a correlation between weather and voter turnout. The paper, published in 2007 by the Journal of Politics, was discussed at the 2008 AMS Broadcast Meteorology conference (audiovisual version here) by Allan Eustis.
Gomez et al. found that every inch of rain above normal correlates significantly 1% reduction in voter turnout. Similarly, every inch of snow correlates significantly to a 0.5% drop in voter turnout.
Eustis pointed out some limitations in this seemingly exhaustive study involving 22,000 weather observations (to resolve weather effects locally) across 14 national presidential elections. For instance, there’s no mention of extreme temperatures or windy weather. Eustis believes extreme weather, not deviations from norms, are more significant in turnout (therefore a linear relationship between precipitation and voting might not be valid).
Cliff Mass of the University of Washington discusses the Gomez paper in his blog and quickly throws a bucket of cold water on the the relevance of those numbers for today’s election, anyway:

Now an inch of rain is quite a bit of precipitation, only occurring during major storms (like Monday in the NW) or in thunderstorm areas.
Furthermore, these results were for presidential elections where people are generally highly engaged and motivated. What about midterm elections like Tuesday’s? If we assume that people would be less excited than for presidential runs would one expect the influence of precipitation to be greater for this election?
And what about the influence of the greater proportion of absentee ballots and of extended balloting times (some places in the U.S. allow voting in the weeks before the election)?

Playing along with the Gomez et al. paper for the moment, however, Mass  predicted (on Sunday) that if Republicans are indeed favored by lower turnout and thus precipitation, the relatively small areas of rain today will have little impact because it will fall areas that are already leaning heavily toward Republicans.
Eustis, however, notes that other studies of weather and elections, unlike Gomez et al., don’t support the adage that Republicans pray for rain (for instance this 1994 paper by Steve Knack of American University).
What Eustis has learned while working for the National Institutes of Standard and Technology, however, is that the weather effects the voting systems, not the voting people: apparently optical scanners can incorrectly process paper ballots, which expand in excessive humidity causing misalignment (see the 2005 SAT scoring controversy).
Ah, for the simple days when gentleman farmers slogged through the mud and rain, got further sloshed with liquor, shouted their preferences to the poll takers, and went home waiting weeks for the results with nary a prognosticating pundit to second-guess them.