I haven't written too much during this election campaign about the polls. But with Election Day upon us, here are some thoughts about the races at the top of Georgia's ballots.
First, how the Senate race has changed over the past month (as of yesterday):
A month ago, David Perdue led by 3.6 points. Now he leads by 3 points. In between, this race got interesting. But maybe not quite as interesting as you think.
At the Real Clear Politics site , you can mouse over this graph to see the candidates' poll averages on any given day. Out of the past 32 days, Michelle Nunn's average was at or above 45 percent on just 12 days. Perdue's number, meanwhile, dipped below 45 percent on just two days. With a pair of non-incumbents here, one takeaway is that someone akin to the proverbial "generic Republican" is not going to get less than 45 percent -- and it remains to be seen if, one the undecideds make up their minds or sit on their hands, how much higher than 45 percent the equivalent Democrat can climb. The evidence to date suggests 45 percent remains something of a ceiling: Over the past few days, Nunn's average actually fell by 1.5 points, while Perdue's number rose by almost the same amount. One of the polls included in the average, by Survey USA , asked about voters who had already cast their ballots. Perdue had a plurality of these voters at 49 percent, and a plurality of those who hadn't voted at 46 percent. Whether the undecideds make a decision or merely sit this one out may be the difference between a runoff and an outright victory for Perdue. Perdue's lead is as large as or larger than the margin of error in three of the past four polls included in the RCP average.
Now, the governor's race , with a longer time horizon at two months:
The matter of incumbency comes into play here, as Nathan Deal has a record to defend. Even before those two months, the governor was sitting below 45 percent for some time; he was even stuck below 44 percent until mid-September. But unlike Nunn, Jason Carter's average has not hit 45 percent a single time during this race. He's only been above 44 percent for eight days, and of late he has dropped back below that level. Like Perdue, Deal has opened up a little bit of breathing room toward the end. Among those who voted early, as polled by Survey USA , Deal won a solid majority. In all six polls used in the RCP average, Deal's lead is larger than the margin of error.
Note that, in both races, the trends have broken sharply in the GOP's favor late in the game.
What does all this tell us about what today's results might look like? A few guesses:
- Of all four candidates, Deal is in the strongest position. Virtually no one would have told you that even two weeks ago.
- A couple of weeks ago, pretty much everyone expected a runoff in the governor's race. As of today, it would be a mild surprise to many of the people with whom I've talked if Deal did not win outright.
- The Libertarian's share of the vote in each race will be no more than 4 percent, and probably closer to 3 percent. In a year where the major-party candidates in both of these races have been on the weak side, the Libertarians haven't been able to take advantage (maybe because one of them actually supports expanding Medicaid, a rather un-Libertarian position). Voting Libertarian as a protest vote will be less attractive with a lengthy runoff in the offing. A lot of people who might have done so will either go ahead and choose a major-party candidate or not cast a vote.
- The Senate race's chances of going to a runoff are 50/50. In favor of a runoff: The black vote looks to be 2 points, maybe 3 points, higher than it was in 2010. Against a runoff: Nunn doesn't seem likely to hit the 30 percent of the white vote she needs to win. She may not even come close: She's at 29 percent in one recent poll , 22 percent in another and 23 percent in a third (the others in the RCP average didn't make their cross-tabs available).