Rich on the Road

In general, people who tell you they don't believe the polls are really just saying they don't like what the polls are telling them.  In this case, if you want John McCain elected president and not Barack Obama -- well, just don't believe the polls.

The fact is that polls over the last 50 years have been very accurate.  In a close race they can't tell you exactly who is going to win but, generally, they are within a couple percentage points of the outcome.

Having said that,  this year IS different.  And, the obstacles facing pollsters are much greater than they've been in the past.  I think voters are much more likely to lie to pollsters (for what reason I can't imagine other than just to be obstinate) and others are much more likely to refuse to answer (which makes more sense to me if they feel it's a privacy issue).  The emergence of cell phones has made it more difficult to get an accurate sample because pollsters have traditionally contacted people on landlines.  But, the overriding problem this year is in figuring out who will actually vote.  Blacks and young people have registered in much larger numbers than in the past.  From what we've seen in early voting, Blacks are clearly turning out in bigger numbers but there is evidence that, once again, young voters are largely no-shows.  Maybe they're waiting for election day.  To get an accurate sample pollsters need to figure out who really is a likely voter.  The old models are not going to work this time. 

Thus, the polls are all over the place.  They all show Obama leading but the range is ridiculous -- from 15 points to a narrow 2 or 3.

I think Exit Polling is going to be even more difficult to do accurately.  Suspect Exit Poll numbers in 2000 caused all sorts of confusion.  In response, everyone was much more careful with the Exit Poll numbers in 2004 but they were worse.  The Washington Post called the 2004 Exit Poll results the most inaccurate ever.   Add to that the growing influence of Early Voting (half the Texas vote or more will be over by today) and Exit Polls get really complicated.

Then there's the so-called Bradley Factor.  Named after a Los Angeles mayor who ran for the Senate decades ago.  Polled as if he were in the lead the whole way but then got beat.  The conclusion was that in Californa then, at least, some white voters would claim they were voting for a Black candidate but then would not actually cast their ballot for him.  Most political observers will tell you that the Bradley Factor has not been seen in decades but....  Well, it's still being tossed around as a possibility this time.

If I was a betting man (I'm not except for the occasional soda wager) I would put my money on Obama.  The polls can occasionally be wrong.  But, they will have to be substantially wrong  for John McCain to win.  Every tossup state will have to go his way and a couple of states that look like sure wins for Obama will have to break right, too.

Still - something tells me unusual things could happen on Tuesday.  John McCain is predicting an upset like we had 60 years ago -- 1948 Democrat Harry Truman -- declared dead in the polls -- instead beat Republican Tom Dewey.

The final poll is always the one that counts.  And, we should know about that one pretty early Tuesday night.  If McCain holds on to Ohio and the other battleground states that George Bush won four years ago, there will be a race.  If Virginia and Ohio -- maybe even North Carolina go for Obama -- it's probably over. 

Either way, I'm ready for it all to be over.

Rich