The New York Times’ John Harwood makes a decent point today–that candidates with leads the size of Barack Obama’s generally don’t squander them in the final three weeks of a presidential campaign.
But this principle is even more iron-clad than Harwood seems to realize. He writes:
Since Gallup began presidential polling in 1936, only one candidate has overcome a deficit that large, and this late, to win the White House: Ronald Reagan, who trailed President Jimmy Carter 47 percent to 39 percent in a survey completed on Oct. 26, 1980.
I’m not sure where Harwood is getting this number from. As far as I know, Carter never led by eight points (or by any statistically significant margin) in any poll conducted in the stretch-run of the ’80 campaign. At best, he was locked in a statistical tie with Reagan; at worst, he trailed considerably.
The Gallup poll that Harwood cites–the one that was completed on October 26, 1980–actually found Carter ahead by just three points, 45 to 42 percent (with John Anderson at nine percent). All other independent polls released around the same time found a similar result:
New York Times/CBS (10-24): Carter 39, Reagan 38
ABC News/Harris (10-26): Reagan 45, Carter 42, Anderson 10
AP/NBC (10-28): Reagan 42, Carter 36, Anderson 10
Sindlinger and Co. (10-27): Carter 38.2, Reagan 37, Anderson 4
Time (10-27): Carter 42, Reagan 41, Anderson 12
Additionally, Gallup conducted a separate poll for Newsweek that was released on October 25, 1980. It found Carter ahead of Reagan 41 to 40 percent, with Anderson at 10. The same poll then pushed "leaners" to choose sides (and excluded Anderson), resulting in a three-point Reagan lead, 42 to 39 percent.
Perhaps the 47 to 39 percent Carter lead that Harwood relies on was the result of "leaners" being included. If so, I can find no record of it–and it would also contradict Gallup’s simultaneous finding (in its poll for Newsweek) that leaners would move to Reagan when pushed.
But even if such a poll existed, it was a clear outlier at the time. Carter never led by eight points; he was lucky to draw even with Reagan throughout October. (In fact, the above data actually represented a surge for Carter, who trailed by wider margins in early October–a surge that compelled Reagan to re-think his refusal to debate Carter without Anderson and to agree to a one-on-one debate a week before the election, on October 28. It was Reagan’s performance in that debate that propelled him to his 44-state landslide on Election Day.)
So while the polls did move dramatically late in the ’80 race, the shift to Reagan wasn’t nearly as profound as Harwood suggests. In his telling, Reagan reversed an eight-point deficit and won by 9.7 points–a swing of nearly 18 points in one week. With that as a precedent, there’s some hope for John McCain, who trails, on average, by about eight points now. In reality, though, Reagan went from even (or actually, very slightly ahead) to a 9.7 win. The very same shift today–the biggest shift ever recorded late in a presidential race–would merely bring McCain into a tie with Obama. Any larger shift toward McCain would be completely without precedent.