A Brief Lesson in Poll Reading
Last week, John McCain was up by about two points in most of the polls. Republicans were ecstatic. The big debate around the conservative universe was what McCain’s margin of victory would be; another hot topic was whether Sarah Palin more resembled George Washington or Ronald Reagan.
This week, Barack Obama is up by about the same margin. Conservatives are depressed. Allahpundit at Hot Air has taken to publishing depressing poll results with listless one word headlines like “crash?”, Jim Geraghty of National Review penned a column reminding voters that an Obama win is “probable.”
It’s no different on the Left. When McCain was up, liberals spent time gloomily pondering the possibility of another “stolen” election. (Liberal don’t lose elections, election are stolen from them). But as Obama rises in the polls, suddenly things are looking up for liberals—the election won’t be stolen this year, and conservatism is staggering through its last days.
For conservatives, the electoral situation went from nirvana to hell in a few days. For liberals, Obama went from dead to ascendant. How could this happen? What did McCain do wrong? And can his campaign recover?
McCain didn’t do anything wrong. The Palin bounce was just a bounce, and what goes up always comes down. Conservatives were stupid for thinking that it would last. It’s September—the two candidates haven’t yet shared a stage, haven’t had to answer any really tough questions, and haven’t faced any noteworthy controversies. A significant proportion of the country isn’t yet paying attention—and that segment will decide the election. (How many politically aware people can possibly be really undecided?).
Polls are useful, but only as very rough indicators of the direction of the political winds. It’s obvious Obama has the momentum right now—but not much. Obama’s lead is well within the margin of error, and any movement at all in the polls towards McCain would erase it. It’s still anyone’s race—either candidate could easily win.
There will be more swings in the coming six weeks. (Does it seem unbelievable that this seemingly endless election is six weeks from completion?) McCain will probably get a bounce from this Friday’s debate—his superiority as a debater is obvious, and Obama probably won’t knock himself out trying to win the first debate. The Palin-Biden debate will represent another McCain bounce—it’s really hard to imagine Joe Biden (who was never a great debater during the primaries) defeating Sarah Palin, whose advantage in the image contest makes up for any substance disparity. (Not that I believe there is a disparity on substance, but if there was, Palin would still do fine).
Obama will get some bounces as well. He’ll probably win at least one debate. Even if he doesn’t, the media will say he did—not so much because of an ideological axe to grind, but because one candidate winning all the debates is boring. He’ll probably make a speech, or McCain will make a gaffe, that will serve to give him another bounce. Conservatives should seek to ride these bounces out without panicking.
There’s only bounce that matters—the one that happens just before the election. It’s at that time that both candidates will attempt to spring their “October Surprises”, that one story that will change the narrative of the election. I can’t imagine what Obama’s Surprise will be; I believe that McCain’s will focus on Obama’s connection to William Ayers. (McCain’s surrogates have mentioned Jeremiah Wright a few times; they have yet to bring up Ayers. That only makes sense—at least to me—if McCain is saving that story for later).
The far too early for the polls to tell us anything about which candidate will win in November. Both candidates will have good weeks and bad weeks, and their respective supporters will get excited, then depressed. But there’s only one poll that matters, and that’s the one on November 4th. Polls before that date have some value—but it’s foolish to get too worked up about them, whether the results are good or bad.
[UPDATE: I think my conclusions about polls are correct, and the past bears this out. But I confess that Rasmussen's recent results (race is tied nationally, McCain leads in Florida and Ohio, McCain trails by just three in Pennsylvania) cheers me. Illogical, yes, but there it is.]