Why the Talking Heads are Still Wrong

Alright folks. I know some have you have gotten a little nervous at the latest polls - specifically the Newsweek and Gallup polls. No fear my fellow Guindibloggers - the results tend to show a Kerry bounce of at most, about 1-2 points. This short-term gain by the senator will, I'm confident, be erased soon enough and will pose more hurdles for the senator in the next month. While all the buzz is that the President seemed tired or annoyed, there were no real gaffes by the discplined Republican. On the other hand, Kerry made a few key mistakes that the Bush team will be pouncing on in the coming weeks. Kerry is the one that made the type of gaffes that result in new questions about his candidacy. In no particular order, here are the four things that we'll be hearing a lot about in the next couple of weeks:

1) the "global test" - this is going to become as famous as "flip-flop" and "voted before it before i voted against it." Plain and simple - this is a huge error.
2) getting allies to join in the great distraction that is the wrong war at the wrong time in the wrong place.
3) giving nuclear fuel to iran (apeasement I).
4) reentering bilateral talks with north korea (appeasement II).

Now, as far as the polls go, the devil is in the details - or in this case, the internals. Let's start with the flash polls that the dems were hyperventilating about after the debate. First, the ABC poll. This poll showed that viewers thought Kerry won the debate by 45-36. However, voter preferences before the debate were 50-46 for Bush, and after the debate wer 51-47 for Bush. So, although people thought Kerry won the debate, it didn't sway any voters. Second, the CBS poll. Again, although Kerry won the debate according to "uncommited" voters, it was actually Bush who gained 12 points as opposed to Kerry's 7 point gain among uncommited voters. Finally, the USA Today poll, which was the most detailed of the three. Kerry won the debate by 53-37, but Bush still maintained an 11-pt edge (down from 14) on who would best handle Iraq, a 10-pt edge on who would best handle the responsibilities of commander-in-chief (down from 13), a 3-pt edge on who agrees with you most on the issues that you care about, was tied in having a good understanding of the issues, was 5-pts more believable, 7-pts more likeable and had 17-pt (!) edge on who is tough enough for the job.

In other words, the average voter was like, "yeah, kerry won. but i'm still voting for bush."

Next, let's take a look at Tradesports.com and the Iowa Markets. Both of these markets had Bush trading at around 65 cents before the debate. Because there was the very real chance that Bush could've put Kerry out of his misery during the debate, one would expect that a tie or even a mild Bush win would result in a slight decrease in Bush's value. In other words, the markets reflected the possibility of a Bush knockout, and thus were probably a bit inflated to begin with. After the debate, Bush's value remained relatively stable until well into Friday morning, when the spin was reaching its peak and the pundits had gone from "yeah, I think Kerry maybe won" from right after to the debate to "Kerry definitely won" 30 minutes after the debate to "Kerry domintated" Friday morning, a classic example of the "pack effect" in the media. Bush's value went down to around 60 cents and settled for a couple of days, which I think reflects a tie or maybe a 1-pt Kerry bounce, to about 59-60 cents now, which illustrates about a 1-2-pt bounce in my estimation. Pay attention to these markets, as my guess is they will be back up over 61-62 cents within the next few days.

Third, the Rasmussen poll. The Rasmussen poll is a 3-day rolling poll whose results are released daily. Its methodical nature is very good in getting a feel for the effect of certain events. In today's poll, the first one that was taken entirely after the debate, Bush maintained the same lead that he had before the debate - 3 pts. This tends to show that there was little, if any, movement due to the debate.

Fourth, the Democracy Corps poll (Carville and Shrum's company) indicates a 2-pt bump for Kerry (Bush still up 2). This is the most detailed poll I've seen.

Next, the Newsweek poll. Remember, this is the poll that showed a 13-pt Bush lead after the convention. Newsweek's volatility can be explained by a few facts. First, Newsweek does not weight the poll for party affiliation. Although party affiliation is not a static demographic, we do know that massive changes in party affiliation over short periods of time are unlikely. In this poll, there were 5% fewer reps polled and 6% more dems polled than the last Newsweek poll. This may be explained by the fact that this poll was taken almost entirely on Friday night and on Saturday in the Pacific and Mountain time zones. Republicans are usually underrepresented in polls taken on Friday nights and Saturdays because Republican men go to high school and college football games at a much greater rate then do dems. Also, these time zones exclude some very red states. Thus, if one weights the poll for party affiliation, this 11% differential explains the difference between the two polls, in and of itself, for all but about 1-pt. Again, this tends to show a Kerry bounce of around a point, if it was weighted like most polls are. This is further evidenced by the fact that this poll gave Bush a 46% approval rating, which is significantly less than the other polls, which further indicates a skewed sample. (Also, Bush's 5-pt lead on the economy turned into a 5-pt deficit although the economy was not mentioned in the last few news cylces). VLWC just showed me the Gallup internals briefly, and it is a similar situation as far as I can tell.

Finally, the polls released today, all taken completely after the debate, suggest that little has changed due to the debate. The ABC/Wash Post poll shows Bush up by 5, last week's was 6. the Pew research poll shows Bush up 5 among likely voters (down from 8), 7 among registered voters (down from 8). The Zogby poll shows no change in the 3-way race (3-pt lead), and a 2-pt bounce for Kerry in the head-to-head (1-pt lead). The CBS poll shows a dead heat, but also shows only a 47% job approval rate (3-6 pts lower than the other polls) and a Bush lead among men of only 3 points (that is Bob Dole territory - usually republicans, as has Bush this year, have double digit leads among men), which indicates a sample similar to the Newsweek and Gallup polls.

To recap, it appears that most polls show about a 1-2 pt bounce, some indicate no change, and others indicate wild changes which upon mere cursory review, are easily discounted and in fact their internals show a 1-2pt bounce. Track the Iowa markets and Tradesports.com. I think Bush has been trading up a little bit over the day as the new polls have come out and the weekend polls have been dissected.

So, there you have it. The talking heads were wrong. There was a slight short-term bounce, that I think will turn into a net loss. Kerry gave Bush at least 4 gifts that he will use over and over and over. Mark my words - Kerry is going to have to explain "global test," leading allies into a great diversion, no nuclear bunker-busters, giving Iran nuclear fuel and appeasing North Korea. (wait, that's 5!)

UPDATE: I totally forgot to mention the LA Times poll. I can't find the link now but this poll was taken for the sole purpose of measuring the impact of the election. Emotional dems have been touting this poll, which shows Kerry up by 1, but this poll was not meant to represent the electorate. It showed Kerry up 1 before the debate and up 1 after the debate, while both candidates picked up 1pt from the undecides. It concluded that the debate had no effect.

Also, I forgot to mention that the Rasmussen poll found that 6% of the voters changed their minds because of the debate, net gain of 3 to Kerry and 2 to Bush, which shows about a 1-pt bounce.

Also, VLWC is talking about "trends." The spin has already worn off with the latest polls today, and tomorrow's bitchslapping by Cheney will help to stop any momentum, even though there really is very little to begin with. Indeed, Bush's expectations are so low for this Friday's debate, that Bush will surely gain from it.