Talk Elections

From Gallup, 9.7.04:

"Is this a significant bounce?

No. To measure bounces in historical perspective, we look at registered voters (because data were only available for registered voters in the summer of election years prior to 1996). This year, Bush only increased his percentage of the vote among registered voters (as opposed to likely voters) by two points. To be specific, before the Democratic convention, Kerry led Bush among registered voters, 48% to 47% on the two-way ballot, and Bush now leads Kerry by one point on that ballot, 49% to 48%.

The average post-convention "bounce" among registered voters historically has been about five to seven points. So the bounces from both conventions this year have been significantly diminished from the historical pattern.

I don't believe that the fundamentals of the presidential race changed dramatically as a result of the Republican convention. Bush was already leading among likely voters prior to the convention. He has simply edged slightly farther ahead. "

I suppose the momentum that has been gained is what's important (and Kerry's dwindling favorability numbers). They also said that candidates in Kerry's situation have won something like 2 or 3 elections since 1948. I'll try to find that one.

From Gallup:

"Al Gore was ahead of George W. Bush by three points right after Labor Day in 2000 and the race ended up being a statistical tie (although Gore received about 500,000 more votes than Bush). Jimmy Carter was 4 points ahead of Ronald Reagan after Labor Day in 1980, but Reagan went on to win by 10 points -- a 14-point swing in the margin between candidates. Carter was also ahead of Gerald Ford by 11 points at Labor Day in 1976, but ended up winning by only 2 points -- a 9-point margin change. Richard Nixon was ahead of Hubert Humphrey by 12 points at Labor Day in 1968, and the popular vote in that race ended up being only a tenth of a percentage point in Nixon's favor -- a 12-point margin swing.

In short, the eight-point margin swing it would take for Kerry to edge out Bush in the popular vote on Election Day is historically well within the realm of possibility"

If I stuck with post-Watergate races (including youth suffrage and campaign-finance reforms), then I should drop 1968 as perhaps demograpically out-of-date. That leaves 1976 R+9 after Labor Day, 1980 R+14 after Labor Day, and 2000 R+3 after Labor Day. As a sample, this is not a set here that looks favorably on Democrats.

Reagan was down by 9 as late as mid October of 1980.

Gee, no Bounce?  What has all the other pollsters been reporting on then?

There's definately been some bounce. Bush seems to be leading by c.6% in three-way race and 4% head-to-head. It's not too different Kerry's bounce following DNC, which steadily eroded during August.

State-by-state polling suggests a pretty tight contest not reflected in the national polls.

Dave