Wednesday, September 3, 2008

As I write this, Alaska Governor Sarah Palin has just finished speaking to the Republican convention.

I believe that anyone who watched this speech, as I just did – with an open mind, will come away with at least four impressions:

(1) Sarah Palin articulately and authentically personifies a populist version of republicanism, the very antithesis of the “country club elitist” stereotype;

(2) If her populist “values & common sense” conservatism ever takes root in the GOP, something like the Reagan revolution will surely follow;

(3) Senator McCain has made an astute choice in a running mate;

(4) Governor Palin will someday be a formidable candidate for POTUS.

For a good account of the affair: Go to this CNN/Time Link: http://www.time.com/time/politics/article/0,8599,1838553,00.html?cnn=yes

And watch this BBC video clip: http://news.bbc.co.uk/2/hi/americas/7597218.stm



The polling industry is never quite spot on. Why is this?

The Built in Error Factor

Part of the error factor is statistical. After all, no opinion sample is quite large enough, quite representative enough and no set of answers is quite honest enough.

The industry standard fudge factor is a 3% margin of error – either way – which is why the presidential race has been in a statistical tie in recent weeks, even when Obama seemed slightly ahead for most of the period.

The “Bradley Effect”

This term refers to a phenomenon well known by both candidates’ internal pollsters. Tom Bradley was a democrat’s dream candidate for California governor: He had been an LADP police Lieutenant who became a lawyer and served five terms as the LA mayor. He was a moderate. He was a handsome guy, in a “good cop” central casting sense. And he was black. He was nominated and ran for the California governorship twice — 1982 and 1986. He ahead in the polls in 1982 against republican George Deukmejian and expected to win. But Bradley lost, in spite of the polls. He ran and lost again to Governor Deukmejian in 1986.

Pollsters are convinced that Bradley’s measured popularity was exaggerated because many of the white voters shaded their answers so as not to appear racist, hence the “Bradley Effect.”

Will the Bradley Effect apply to Senator Obama, the “post racial” candidate? Frankly, no one knows. My own opinion is that the Bradley Effect is still real, but that it has been weakened by declining racism and Obama’s “Tiger Woods” persona — after all, we’re no longer in the 80’s.

But I still estimate the 2008 Bradley Effect about 2%, This is not because we are still a rabidly racist country, but because there is a second factor operating. Note that the country as a whole is considerably center/right of the post FDR leftists who now run the democratic party. Because the liberal-left dominates the culture, many voters feel that it is not “cool” to admit to “being uncomfortable” with Barack Obama’s left wing associations and positions, lest one be shunned as a racist. We Americans still don’t level with the pollsters.

The Finish Line Effect

Even exit polls aren’t perfect. Those who talk to pollsters after just having voted are more likely to be partisans. The non partisans are less inclined to talk. Recall President Bush’s win against John Kerry where network pollsters were misled during the voting to expect a democratic win.

The overriding problem is that a significant block of voters can and do make their final decision, especially in a close race, at the last possible minute. This is particularly true when a voter is leaning to a candidate nominated by his or her same party, but harbors misgivings.

Polls show that roughly the same number of voters in each party are firmly behind their nominee. This means that a significant number of voters in each party are up for grabs. Right now it appears that about 40% of the voters are essentially committed to Obama and 40% to McCain, plus or minus 2%. Yes, this is a crude estimate, but it means that the election will be decided by about 20% of the voters, roughly half of whom will make up their minds in the last few days, many even on the last hours, and some only in the privacy of the polling booth at the last minute.

The “Turn” Factors

Knowing that the Finish Line Effect is real, can we identify the factors that will tend to “turn” the undecided voters toward the end? The problem is captured in two questions:

[1] What is each candidate’s “personal revelation” trend?

Put another way: Which candidate benefits (or loses) more by the drip, drip, drip of personal revelations about his or her character, biography, accomplishments and positions? This boils down to the impact of hidden negatives and hidden positives in which a pattern of revelation begins to favor one candidate over the other along a timeline leading to the pre-election weekend.

The operatives of both campaigns tend to secretly agree that the “revelation” trend will favor McCain, whose dramatic personal history is a character parable and whose long tenure mitigates against last minute negative revelations.

But Obama’s remarkably candid autobiography presents an interesting dilemma, because it was so well written, and because he is so well liked on a personal level. Will he benefit from a “likeability Teflon” factor that protected someone like Ronald Reagan from scandal? I think that the Illinois senator is probably insulated from traditional scandal, but not from the impact of those actions, positions and associations that position him well to the left of the mainstream. Ideologically, Obama’s history is a “target rich” environment.

[2] What is the candidate’s “situation match” trend?

Stating this another way, how do a candidate’s capabilities and character match up with a perceived “crisis” trend?

For example, the more voters know about McCain and Obama, the more the McCain is trusted in a major international crisis. Will we be trending into an international crisis in mid October? Does the sun rise?

For example, the more a voter laments the “sorry state of the country”, the candidate most strongly positioned as an effective critic of the current administration is favored. Conventional wisdom favors Barack. Will we be worried about the economy in October? Does the sun set?

I suspect that the election may be decided on voters’ trust and confidence in the candidate’s competence. The question of perceived competence can be decisive, especially when a voter’s “things are bad” assessment mutates into “things are scary.” Depending on the level of voter anxiety, trust may trump hopeful idealism.

The Preference for Divided Government

Bill Clinton did his best work when he was forced to deal with a Republican Congress. Reagan was a more popular president when people knew that a Democratic Congress could hold him in check. In my opinion, Obama would run better if the Republicans controlled the congress and McCain will run stronger (especially at the finish line) because the Democratic Party controls both Senate and House.

Bottom Line

There is always movement in the last week or so of the campaign. Sometimes it merely cuts into an large lead, sometimes it reverses the outcome.

I am now willing to make three predictions:

(1) Unless and until Obama gets a durable breakout lead (something exceeding 5%) within a couple weeks following the republican convention, this race will be a toss up in early October.

(2) If the candidates are very close (within a couple of points of each other) in the last week of October, Obama will almost certainly lose.

(3) Even if Obama is ahead by only 4 points going into the last few days of the election, Obama will lose unless somehow he is gaining momentum at that point.

I make these predictions on the basis that Obama probably has peaked. If that assessment is true, time is his enemy.

Mark your calendars and stay tuned.


Leave a Reply