This is the spreadsheet I made after the Iowa and New Hampshire results were in. It is weighted based on:
1. 18-35 year old male population (per capita)
2. RonPaul2008.com donors (per capita)
3. State score on PledgeForPaul.com
The 2 items I would also like to add (but don’t have the data for) are meetup group sizes and libertarian party members. After plugging in the South Carolina and Michigan data, I think this model is holding up pretty well. South Carolina had less votes then expected while Michigan had more then expected. The largest error so far is Nevada, which had a 3.5% higher then expected turnout. Las Vegas is the #7 biggest group on meetup, so taking into account meetup group sizes may have put this projection more in line with reality.
EDIT - There was an error in the per capita donor data feed… I updated all of these values (today is a money bomb day and aside from the data feed being reversed, these per capita donation levels significantly changed in 5 hours). Here is the revision… I’ll update again after Florida:
Excel Sheet: http://www.mediafire.com/?dzrzdzxv4jt
