Nimh makes an excellent point about the often overlooked impact of plain old prosaic advertising on poll numbers.

AP

While I agree that advertising dollars are important (and adore the graphs!), the post got me thinking about some of the other factors involved in Obama’s surge. Advertising is an underestimated piece of the puzzle, but still just one piece of the puzzle. So here are some of the other elements that I think are at play:

Obama’s 50-State Strategy

This has a lot to do with the Obama campaign’s relatively large advertising budget — but it’s not just about advertising. Obama’s been spreading McCain very thin in many different ways, as McCain has to spend time and resources defending red states, rather than being able to focus on battleground states. The thinner things are spread, the less McCain is able to campaign effectively (not just advertising but field offices, rallies, paid staff, etc.).

The University of Wisconsin Advertising Project “codes and analyzes nearly all of the political advertising that is aired in 2008 federal and gubernatorial races across the country.” Yesterday it released a very interesting report on the two presidential candidates’ advertising in the week of September 28-October 4 (h/t Marc Ambinder). There’s a bunch of goodies in there, data-wise.

$28 million in one week

First of all, there’s the sheer volume of advertising that’s going on. Baffling amounts of money are being spent on equally stunning numbers of ads. In that one week alone, the two campaigns spent over $28 million on TV advertising.

That’s almost twice as much as in the first week of September. It’s also one and a half times as much as “the Bush and Kerry campaigns and their party and interest group allies spent” in the equivalent week of 2004. (Remember the reports back then about the unprecedented role money played in a record-breaking year of campaign spending?)

$28 million in one week. I mean, you could have 28 million young Africans immunised against meningitis for that. Just saying.

The result was that in the Las Vegas media market, Obama ads were aired 1,288 times in one week, and McCain ads 712 times. That’s a lot of ads.

Charting the ads

Secondly, the sheer extent to which Obama is outspending McCain on the airwaves. And the revealing differences in where they spend their money. In this one week, “the Obama campaign spent just under $17.5 million while the McCain campaign and the RNC spent just under $11 million combined.” I’ve graphed it, of course. This is by how much Obama is outspending McCain – and where:

On a total aside in a post on McCain’s (lack of) campaign strategy, Daniel Nichanian at Campaign Diaries has this remark, between parentheses, about Sarah Palin’s recent telephone interview with Bill Kristol:

(Kristol acknowledged that it sounded like Palin was being coached by staffers while on the phone with him)

No way. She had staffers sitting in on a phone interview with a sycophantic columnist to make sure she didnt goof even in that setting?

But it’s true: in his complete softball of an interview, Kristol playfully asked Palin whether, “since she seemed to have enjoyed the debate, [..] she’d like to take this opportunity to challenge Joe Biden to another one.” Silence followed: “There was a pause, and I thought I heard some staff murmuring in the background (we were on speaker phones).” Eventually she passed on the notion of a challenge.

Now I’m a layman, I’ve never accompanied any candidate on his or her business — is this normal?

God knows what got in the daily tracking polls today, because they’re all out of sorts.

Daily tracking polls update, 10/8: click to enlarge

Yesterday, the Diageo/Hotline poll suddenly had Obama’s lead drop by four points, from +6 to +2. That’s an unusually steep drop for a tracking poll. It was the biggest day-on-day change in any of the daily tracking polls since 7 September, over a month ago. It was all the stranger because none of the other tracking polls showed something similar: Obama’s lead dropped by one in the R2000/Kos poll, increased by a point in the Gallup poll, and was unchanged in Rasmussen’s.

Today, the weirdness continues. The Hotline poll has Obama’s lead down another point to +1. Rasmussen has his lead dropping by two points as well, from +8 to +6. But Gallup has it storming ever upward, today from +9 to +11; it has Obama at 52% of the vote. That result, as Gallup’s Jeff Jones points out, is “the best for Obama during the campaign, both in terms of his share of the vote and the size of his lead over McCain.”

So which is it? Is Obama up by 11%, 6% or 1%? That’s quite the difference. And is he moving up or down?

Earlier today, I wrote “Can we Wordle? Yes, we can!” Wordle being a toy for generating “word clouds” from any given text. I created a Wordle of all of Obama’s answers during last night’s debate. (Again, the hat-tip for the idea goes to The Monkey Cage, which did a Wordle of the Vice-Presidential debate, both candidates in one).

Still using the of same transcript, here’s the Wordle for John McCain’s words:

Click to see large.

So how do the two compare? And what words stand out for prominence – or absence?

Process-wise, what struck me is that Obama spoke slightly more (some 6,720 words) than McCain (about 6,550), and that McCain made more short quips and interjections. The dynamic of frontrunner versus underdog? McCain needed to be on the attack after all, push for some kind of break to revive his chances, while Obama’s main goal must have been to keep ‘er steady and not disrupt his comfortable lead in the polls.

An easily recognizable word in the McCain Wordle is friends – as in, “my friends”, McCain’s catchphrase. Another marked feature of the McCain Wordle is that America, Americans and to a lesser extent American are among the most prominent words. United and States are big too. The American President America is waiting for! In the Obama Wordle, America, American and Americans are all present, but very small.

The prominence of Well, like and look in McCain’s Wordle makes him look curiously like a Valley girl. But he’s also got thank fairly large in there, and Tom (Brokaw). Obama had less time for such niceties and spoke more directly to the audience, over Tom’s head.

Can we Wordle?

Yes, we can!

Wordle is a toy for generating “word clouds” from text that you provide. The clouds give greater prominence to words that appear more frequently in the source text. You can tweak your clouds with different fonts, layouts, and color schemes.

Well, here goes: this is the Wordle of all of Obama’s answers during last night’s debate, on the basis of this transcript.

Click to see large.

A hat-tip is in order: I got the idea when I saw the Wordle on The Monkey Cage of the Vice-Presidential debate. That Wordle was of everything both candidates said though, which seemed less revealing to me than what you’d get Wordling each candidate’s words separately.

I’ll do one for McCain later today..

I adore politics. My job is politics-related. I’ve been watching this election cycle closely from the very beginning.

So it’s kind of hard for me to remember that there are a whole lot of people out there who just don’t care very much about politics. Now is when they start to pay attention.

And Barack Obama has shown over and over again that when people pay closer attention to him, he benefits (and his opponent suffers).

Take these three graphs from Pollster, indicating voter preferences in the lead-up to the Democratic primaries in Ohio, Wisconsin and Missouri.

Ohio

Wisconsin

Missouri

See how that orange Obama line goes up-up-up, but especially in the very last stretch before the primary?

Tracking polls update, 10/6 (click to enlarge)

Quickpost this image to Myspace, Digg, Facebook, and others!

The daily tracking polls haven’t shown a great deal of movement over the past three days. As you can see (at least if you click the graph and look at it in original size), the Research 2000/Kos poll is still off on its own, measuring a 12-point blowout lead for Obama/Biden today, but it’s been stable. The three other daily tracking polls are roughly in agreement: Gallup and Rasmussen have an 8-point Obama lead, the Hotline poll a slightly smaller 6-point lead. That’s all roughly the same as the last three days, when all those three polls had the lead at 6-8 point too.

On the other hand, this stability in the daily polls does confirm the slight move that I signaled in the last polling update three days ago. I wrote then that “after five days of similar numbers,” with an average Obama lead of 6.3-7% on each, there was “a hint of new upward movement”, with the average lead moving up to 7.8%. Well, the Democratic ticket has kept that extra point and another extra half: the average Obama lead has been 8.3-8.5% since.

This seems to have become something of a pattern in the past month or so. Obama moves up a couple of points, stabilises at the new level for a few days, then moves up a couple of points again.

The Research 2000/Kos poll, which has had the largest or shared largest Obama lead in all but 3 of its 26 days of existence so far, is becoming a bit of a distraction. It’s not entirely easy to take a poll at face value that so consistently diverges from the others, especially if it’s a poll commissioned by the Daily Kos and it’s diverging in favour of Obama. But even without taking the R2000/Kos poll into account, today’s numbers are very good for Obama.

The daily tracking polls have been relatively stable over the last three days, since the previous polling update here; hence also the absence of intermittent updates. But today’s polls are interesting because they do not, or hardly, include any impact the Veep debate might have. So they can be used as a yardstick for measuring any such impact in three days’ time, when most of the tracking polls will have a sample completely from after the debate.

(There’s some trickiness involved, of course, in that the Veep debate will hardly be the only development impacting opinions; what about the Senate vote for the bailout bill, for example? Or, as Gallup points out, “the new Labor Department report out today, showing a bigger job loss in September than many analysts had predicted”?)

Daily tracking polls, 10/3 (click to enlarge)

Here’s the graph. Gallup’s poll bounced around a little the last few days, with Obama’s lead shrinking from 8% to 4% between Monday and Wednesday, and increasing to 7% again by today. The Rasmussen and Hotline poll have remained almost entirely stable and similar, showing an Obama lead of 5-7%. And the Daily Kos tracking poll, conducted by Research 2000, is off on its lonesome showing a significantly larger Obama lead of 10-11%.

Quickpost this image to Myspace, Digg, Facebook, etc.!

Reviewing the average of the four polls, however, there’s a hint of new movement today.

Seems like an appropriate link to recommend tonight: The Poetry of Sarah Palin. A whimsical, ironical (re)take on Palin’s words, written with a gentle and creative humor. The author, or should I say editor, is Hart Seely. You might remember how he turned Donald Rumsfeld’s cryptical utterings into The Existential Poetry of Donald Rumsfeld.

“You Can’t Blink”

You can’t blink.
You have to be wired
In a way of being
So committed to the mission,

The mission that we’re on,
Reform of this country,
And victory in the war,
You can’t blink.

So I didn’t blink.

Trivia absurdism of the day. At Beyond Red and Blue, Robert David Sullivan last month did some serious research into bellwether states:

I calculated the percentage-point differences between each county’s swing and the nationwide swing for each election from 1980 through 2004, then added them all up to find out the places that have deviated the least from the US total over that time. (For example, there was a swing toward the GOP and George W. Bush of 2.86 points in the last election. A county that swing 12.86 points toward Bush and a county that swung 7.14 points away from Bush would each be penalized 10 points for that election.)

The result is a map, list and Excel spreadsheet of the Top 50 Bellwether Counties, 1980-2004.

Which county is #1 — the single most conformist county in matching the country’s overall swing for the last seven elections?

Defiance, OH.

Today’s daily tracking polls offer a mixed picture, roughly showing a stabilisation of Obama’s lead.

Daily tracking polls, 9/30 - click to enlarge

Gallup has Obama’s lead down 2 points, from +8 to +6, but the other three trackers all have his lead increasing by a percentage point. Rasmussen and Hotline both have it growing from +5 to +6, while Research 2000, the Daily Kos-sponsored poll, has him going up from +9 to +10.

Both Rasmussen and Hotline have basically has Obama’s lead stable at 5-6 points for four days in a row now. Basically, it seems Obama’s lead grew as the financial crisis first became clear, then stabilised, made another jump when McCain “suspended” his campaign, and now has stabilised again.

Obama’s well; impact of debate hard to distinguish from impact of McCain’s political stunts last week

McCain announced the suspension of his campaign on Wednesday the 24th. The first day of polling after this announcement was included in the daily tracking poll releases of the 26th, and that day Obama’s average lead in the tracking polls jumped from 3.3% to 5%. As the three- and four-day rolling polls started including more days from after McCain’s announcement, Obama’s average lead increased further, to 5.5% on Saturday, 6.5% on Sunday and 6.8% on Monday. The public was obviously not impressed by McCain’s political theatre.

By then of course the debate had taken place too, though. The first day of polling after the debate was included in the releases of Sunday the 28th, when Obama’s average lead jumped up a full point to 6.5%. But that jump can’t be read to directly indicate the impact of the debate. That’s also the day, for most of the tracking polls, when the last pre-suspension day of polling rolled out from the sample. In short, just like the previous two days which also saw notable gains for Obama, it was a day on which a day’s worth of polling from before McCain’s suspension of the campaign was replaced by a day’s worth of post-suspension polling.

It’s the two days since that which seem a more logical focus for measuring the impact of the debate. And on these two days, Obama’s lead basically stabilised, with his average lead rising from 6.5% to 6.8% and 7.0%. Just going on these tracking polls, the debate was indeed something of a draw, then. Which arguably counts as a tactical victory for Obama, since he’s the one ahead, and a draw is a perfectly adequate result if you’re 7 points out ahead.