Sunday, August 01, 2004

Convention Bounce, or Flop?

"Mr. DeMille, I'm ready for my convention bounce...." (apologies to Billy Wilder)

Two polls in the last two days suggest that Kerry may be in unique, and uncomfortable, position of going for his convention bounce and just breaking off the diving board.

Now, it's important to remember that what we are talking about here is the "convention bounce." There is no evidence that polls this early, before the second convention even takes place, tell us ANYTHING about the election, which is of course what matters.

Having said that....holy cow! What we have is two polls: Newsweek says Kerry gets a four point bump (though it could be more), and CNN/USA-T/Gallup say Kerry lost a point, but more importantly Bush gained (they have it Bush 50%, Kerry 46%, compared to Kerry 47% and Bush 46% two weeks ago). (Though, as the comments to this post point out, it is important to remember the resident statistician for "Car Talk," Marge Inovera)

One way to read that is that the convention bump was much weaker than normal. The other way to look at it is that Bush is just showing more strength, and it happens to be at the time of the Dem convention, so it all gets mixed up. (See Bloomberg story for more background)

Here is what MSNBC/Newsweek had to say:

Kerry’s four-point “bounce” is the smallest in the history of the NEWSWEEK poll. There are several factors that may have contributed to the limited surge, including the timing of the poll. On Thursday, Kerry had just a two-point lead over Bush (47 percent to 45 percent), suggesting that his Friday night speech had a significant impact. Additionally, Kerry’s decision to announce his vice-presidential choice of John Edwards three weeks before the convention may have blunted the gathering’s impact. And limited coverage by the three major networks also may have hurt Kerry.

My own diagnosis: The lack of TV coverage mattered. The Dems sang a lot of Kum-ba-ya, and talked about how the Repubs were "negative." But you can only get away with that if you are the incumbent. If you are the challenger, you have to attack. Kerry is so sure he is better that he is running as if he were the front-runner. Remember who said this? "This election is not about ideology. It is about competence." If Dukakis was right, the "competence" vote lost him all but a paltry 11 states. The utter self-confidence of the Dems that people will flock to their non-message is surprising, and may kill them again, just like 1988.

In the headliner speech at the convention, Kerry's chowdah had too much butter and not enough clams. He failed to accomplish his key task: project strength in foreign policy. We all know that Kerry will improve U.S. relations with other nations. The question is whether Kerry can be trusted to DEFY other nations if U.S. interests demanded it. The answer, in some key voters' minds, seems to be "no."

3 comments:

Anthony said...

There's also the fact that unlike previous elections, a much smaller pecentage of the electorate entered the convention season as "Undecided." Given the smaller group, Kerry would have to start converting solid Republican votes to get a bounce even near to the proportions of the 1990s and 1980s.

Matt Dowd predicted a 15-point bounce, which would entail capturing every undecided voter, as well as a 5% block of Republicans! Now THAT is playing the expectations game. Just wait for the depiction of John Kerry as a "International Debating Champion" in the weeks preceeding the debates...

Let's also remember the insignificance of national polls. Bush garnering 80% of Texas means little to the folks in the Electoral College. The only polls that matter are those taken in October and that sample likely voters in Florida, Pennsylvania, Ohio, Missouri, New Mexico, etc.

Anonymous said...

Killer Grease,

What ever the reason may be, I think that you need to address whether the bounce and spread are statistically significant. If the margin of error is +/- 3% then there is effective no discernable boune. Likewise, there is no statistically measurable difference between Bush and Kerry before or after the convention.

Mungowitz said...

I'm skeptical of the whole idea of a "bounce."

As I've said elsewhere, it is not really an intercept shift at all. More like plucking a guitar string. Once the oscillations damp down, you are pretty much back where you started.

and....the stat signif point is a good one, but see above. if there is no such phenomenon as a "convention bounce", it is not surprising that we have trouble measuring it.