Thinking about polling and prediction

October 28, 2008

I have a few theories hunches on polling this cycle…and here’s how I see the final electoral map (Montana is the only true toss-up in my opinion)  OBAMA/BIDEN=326; MCCAIN/PALIN=209; UNKNOWN=3

Myles' final prediction for election night

Myles' final prediction for election night

1.)  Obama is overperforming on some polls because the people who support him are proud to say “Yes, I’m voting for Obama”; McCain supporters probably are less likely to respond at all to pollsters

2.)  Obama is underperforming on some polls because they don’t take in account cell-only voters. I have a detailed post on this subject on my personal blog.

3.)  Obama is underperforming on some polls because they don’t contact newly registered voters and there are numerous data points that indicate the Obama campaign and Democrats have registered millions of voters across the country.

4.)  The “Undecideds” are mostly McCain supporters who don’t want to admit they support McCain.  Therefore, McCain’s numbers will close on election day.

5.)  Obama will win any state he is currently up by +5% (and polling over 50%)

6.)  Mason-Dixon is given too much credit for accurate polling

7.)  Anytime a pundit says, “A democrat hasn’t won here since…” expect Obama to win that state.

Make your own map at Real Clear Politics.

Advertisements