Obama leads by 1.74%

Obama is ahead, but Clinton and Edwards are all within striking distance of each other and the latter two will benefit most from the caucus rules

Incorporating the Selzer poll (which had its end date before the ARG poll) I’ve come up with the following projections for the Democratic contest in Iowa: Obama 27 Clinton 25.26 Edwards 23.07 Biden 7.61 and Richardson 4.53. Hillary’s share of the vote is falling slightly, Obama’s is rising and Edwards seems to have pulled back from the 15% that he was predicted in late October. At the bottom of the scale Biden seems to be making a late increase, although it is clearly too little and too late, Richardson is falling of the charts and Dodd is doing so badly that betting on him withdrawing from the contest might be a good idea (and I haven’t even tried to estimate his vote share on the chart).

Using correlation analysis to see if there are any voting blocs (negative correlation indicates that two candidates are fishing in the same poll), I predict that on Caucus night frustrated Biden voters will switch to Obama, Richardson voters will switch to Clinton, Edwards voters will switch to Clinton, Obama voters will switch to Clinton, Clinton voters will switch evenely between Clinton and Obama and undecided voters will switch to Edwards. Assuming 50% of Biden and Richardson’s and 25% of Clinton’s, Obama’s and Edward’s votes will be wasted due to the 15% hurdle in each precinct, I project that Clinton will win with 32.60%, Edwards will get 29.83% and Obama will get 27.21%.