The results in Iowa will tell us about whether robo-call polls or other traditional polling methods

[Attorney in Texas]

are more reliable.

538 and other polling analysts have confirmed that the robo-call polls are projecting different results than the many polls conducted according to all of other polling methods.

The Democratic race in Iowa is an excellent example of this phenomenon.

If you look at just the robo-call polling, you see Clinton well ahead:



However, if you look at ALL of the polling except the robo-call polls (the live cell + landline phone polls and the internet-based polls come to similar results), you see a race that is a dead heat with Sanders rising while Clinton falls:



The reason why the poll aggregation websites (538, RCP, Pollster) show Clinton with a slight lead is that they average the robo-call polls (which consistently favor Trump and Clinton) with the other polls which use a traditional methodology (so Clinton's lead is smaller than her robo-lead but the race is not tied as the non-robo-polls indicate).

Check back Monday night to see which polls were more accurate!

[CoffeeCat]

1. If you want incredible polling accuracy

Ann Selzer is releasing the last Iowa Poll before the caucuses, tomorrow (Saturday).

She predicted Obama's 08 win as well as other races that other pollsters didn't accurately predict.

Her polling methods (especially when it comes to the Iowa caucuses) are impeccable. Nate Silver gave her an A+ rating, which he has only give to three pollsters.

Should be interesting.

The Saturday-before-the-caucuses poll in 08 was nearly spot on with the final result.

It's kind of nerve wracking waiting for this one.

[Attorney in Texas]

2. I agree. Selzer uses traditional live cell + landline polling. Her results will almost certainly

show Sanders and Clinton within the margin of error from each other. This will be a cliffhanger.