Iowa Overview

1. First I want to acknowledge my mistake of assigning a center percentage of 3 to Christie when I ranked him behind bush.  I was struggling to get the magic 100.00% and had some difficulties.  My ranking was almost perfect only switching Kaisch and Huckabee. I did predict that Marco Rubio would out-perform his poll numbers but I did not expect him to do as well as he did.  Looking forward to next tuesday, I am going to do an historical bayesian adjustment of the New Hampshire polls and only include very recent polls.  Let's breakdown of what  I learned yesterday:

1. Ground Game Matters

A strong organization can beat or almost beat a celebrity candidate. But celebrity and popularity can still help.


2.  Trump did increase turnout (of non-Trump Supporters)

Since I am a republican and therefore think like a republican and if I normally didn't caucus but if there was a threat of a candidate that was in my mind disruptive to my party I would definitely vote.

3.  Watch the highly educated or higher income voters. (I already figured this was important)

In the very few polls  available with fully demographically stratified results showed a much similar picture to what actually happened if you adjusted for the average income level or education level of an Iowa voter.

4.  A lot can happen in 72 hours.

The final Des Moines Register Poll ended about 72 hours before caucus day.  It showed the decline of Bush and the rise of Rubio, and the lower trump lead (it was in the margin of error).  Momentum carried those changes and allowed them to show up on Causus night.

5.  Thinking like a Republican Matters

One of the unique perspectives I carry to my predictions of the GOP is because I understand the mind of the Republican matters.  I think the instinct I had that more moderate (or at least not Trump Supporters)  voters would turn out fearful of the polling data.  Regardless of your political beliefs your model and your predictions should reflect the minds of who you are predicting.

What I Think Will Happen In New Hampshire

Given what happened last night, I need to rethink the New Hampshire process.  Since Trump was polling with a double-digit lead on Rubio but lost by 2% , maybe his data needs to be treated with a grain of salt and the undecided voters (who i predicted would go to Cruz or Rubio) should not be proportionally reassigned but used as  a  method to correct the model.


What Last Night Means to the Nomination Process

Although  Trump or Rubio (and maybe Cruz) maybe incredibly close to the nomination a brokered convention is still a possibility. And the unbound delegates in the nomination may end up with the final call, and if they are looking for a win in November they will probably pick Rubio and a more Tea Party partner..