Potentially In McCain's favor-

The Bradley Effect-

Ken and every other Republican have long been celebrating that America's inherent racism will cost Obama the election.  The idea is a voter, not wanting to appear racist, will lie to a pollster and pretend he is voting for the black guy.  Once in the booth, he will revert to his Klansmen true self and Obama will lose.

The history here is the L.A. mayoral race where African-American Bradley underperformed his polling in the final count.  A similar, but far less dramatic and clear cut event was the Gubernatorial campaign of Douglas Wilder in Va.

Many are now convinced any black candidate has a built in -(insert number) in the polls.  Ken usually sets it at 8 or so.

MY TAKE- It's an interesting argument that appeals to our worst fears about ourselves.  The problem is its not really about us.  The Bradley race was in 1982!  The Wilder race (which is very unclear as to this effect anyway) was in 1989.

I suppose I believe we are a completely different society than we were 26 years ago.  My God, 1982 was the year McCain was sent to Washington.  To project those voters behaviors on race to today seems a stretch.

Here are some numbers to think about.  In 1983, polling showed HALF of all Americans were opposed to mixed race marriage.  Two years ago that number had fallen to 15%.

In 2000, pollsters asked if America was ready for a black president.   38% said yes.  Now? 68%.  That is a big change in 8 years.

But that is not all.  If you read the wiki entry I linked above on the Bradley Effect, you will see since the early90s, it seems to have disappeared. 

To sum up, I see the people pimping the Bradley Effect as why Obama can't win as I would the people who insisted JFK was unelectable as a Catholic because Al Smith lost in 1928.

Cheatin’

There is a moment that happens every election that drives me crazy every time.  I know its coming but it doesn’t help.

It goes like this.  The news comes sometime before noon and begins a story like this.  “Both Republicans and Democrats look for every edge in elections but sometimes they go to far…”  This is invariably followed by some slimy story about Republicans telling minorities the election has been moved or that their Ids will be run for warrants or that we’ve reinstituted the poll tax.  It is never followed by a story about Democratic cheating.  We are just assured it happens.

But I digress.

The GOP will cheat.  They will cage.  They will lie and intimidate to drive down voter turnout.  They might even use rigged voting machines. 

MY TAKE- There are two conflicting forces that make me wonder how intense these efforts will be. 

The first is the current administrations malfeasance.  I firmly believe that Cheney and his cronies were willing to do ANYTHING in 2004 to ensure they won the White House.  The driving force being that the place was a giant crime scene overflowing with evidence of the lying and manipulating of the first term.  It seems clear to me the Bush folks are terrified of what will happen if someone who is not friendly to them comes in and starts looking through the files.   This suggests that current officials will be willing to do whatever needs to be done to keep Obama out simply for the sake of their own self-preservation.

On the other hand, the GOP base seems fairly indifferent to McCain.  Since I imagine most cheating is done by volunteers on a local level this may be the stronger force.  Some people may just not feel as motivated to get up early on a Tuesday and go try to scare the shit out of or confuse little old black ladies all day for McCain as they were for Bush.  And before you go there, I do not think Palin, in her current state of approval, is enough to turn the tide on this.  All in all, I expect LESS cheating from the GOP.  I am more worried about one big event that the Cheney Administration could orchestrate than I am individual, foot soldier level cheating.

Potentially in Obama’s Favor

Modeling

The things that could potentially help Obama relate to modeling.  Modeling works like this:

Pollsters collect raw data from some number of potential voters.  They also collect demographic information.  They then adjust the percentages the data suggest to match their predictions of how the actual electorate will look.

Example:  Let’s say Gallup calls 500 people for a tracking poll.  Let’s say 30 of those people are between 18 and 25 (6%).  Now assume Gallup is, using past elections as a guide, assuming 14% of voters on Nov 4^th will be in this age group.  They then weight the 6% of actual respondents to make up 14% of the poll “results”.  ; ; They do the same thing with party affiliation, so if they expect 36% of voters to identify as Dem and call 500 people and 42% identify as Dem, they downgrade the responses of those 42% to make 36% of the “results.”  ; ; This election is a nightmare for modelers.

The Ground Game- As a football fan and coach, I can promise you if this were being played on the gridiron, McCain would have no chance.  Obama’s ground game is so much stronger there is nothing McCain could do to overcome it.

In politics, its much less clear.  

My Take- Obama’s ground game has the potential to get a much higher percentage of his supporters into a voting booth than Dems have in the past.  Since all polls model their reported results on what happened in previous elections, this could create a serious discrepancy in Obama’s favor.

Going back to my discussion of modeling, this means that if in the past, say 80% of people who claim they will vote Gore or Kerry showed up to vote, the polls assume 80% will show up to vote Obama.  The problem this time around is that voter enthusiasm, registration and participation on the Dem side is dwarfing anything we have ever seen since these methods were adopted.
So each pollster is left asking how much of this is real and how much is an illusion.  Rasmussen, for example, seems to almost entirely discount this factor, assuming the model for this election will be almost identical to the last two.  This is one reason why Rasmussen has skewed more McCain that most other polls.  The Battleground poll is assuming there will be no change at all in black voter turnout.  This and two other key problems help to explain why it is getting such different numbers.  More on that later.  I am not predicting who is right and who is wrong here, I am just pointing out one possible cause of error.

Those Pesky Kids-

Similarly, there is great debate on how to handle the youth vote.  Battleground is assuming the percent of votes cast by those under 30 will be the same as in 2000 and 2004.  Other poll services are desperately trying to guess how high to go without going too high.

My Take- I think the kids will vote in higher numbers than ever before.  The reason many pundits insist they won’t is 1968.  But ’68 and this election have almost nothing in common short of young voters interest in the primaries. 

See, back in ’68, a whole lot of younger voters got involved in the primary because they weren’t too fond of the idea of the government forcing them to go halfway around the world to die in a swamp.  So they got involved and supported the anti-war Democrat, Bobby Kennedy.  They voted like crazy in the primary and helped Bobby to a lead in the nomination race.  Then Bobby got shot.  The Dems turned to a pro-war candidate and the kids decided they really had no reason to support this new guy since he disagreed with them on the only issue they cared about.  Consequently, given a choice between a racist, vague on the war Nixon and a pro-war Humphrey, they stayed home.  Forty years later, everyone in the political world assumes we are having a replay of that.  I don’t see it.

This time there is no issue dividing the Democratic party, there were personalities dividing it but no issues.  More importantly, the kids candidate WON.  Why the hell would they abandon the guy they got nominated.  If you want to talk about 08 in terms of 68, the only relevant question is would the kids have voted for RFK if he had lived.

Now to be fair, no one knows the answer, but I would move this discussion from something we pretend to have data on to an open question.  I am giving the kids the benefit of the doubt.  This is a Civic minded generation, as others have been in the past.  Politics are in fashion and the Democratic candidate has embraced, and been embraced by, the youngsters.  I suspect any polling company not modeling a slight to reasonable rise in young voters if going to skew McCain.  We shall see.

Cell Phones-  Another modeling nightmare is cell phones.  Not only are the numbers hard to get but younger people just won’t answer numbers they don’t know.  Consequently, cell phones are not really getting polled well.  Rasmussen and Battleground just won’t poll them at all.  Other agencies are trying to use age group adjustments to project a “what if these damn kids actually answered their phones” scenario. 

My Take- What this means is Battleground, Rasmussen and any other poll who won’t call cell phones show about a 2 point skew (according to 538.com) to McCain.  Others are kind of guessing.  Either way it injects uncertainty into the data.