PROBABLE ROMNEY CO +2.2 (21 surveys, MOE 0.9) FL +3.3 (26 surveys, MOE 0.8)As you can see, even though - excepting Iowa and Wisconsin - the margins are very small, because of the pooling of numerous surveys, the margins of error are similarly very small. There is no problem with sample size. The problem is uncertainty concerning the assumed turnout model. This uncertainty will only be resolved upon a tabulation of the vote. In considering the reasonableness of my assumptions, I have considered the following information: The best economic model available (Ray Fair's) predicts a 2 point victory for Romney in the nationwide popular vote. My estimate for Romney's margin in the nationwide popular vote, based on 59 national polls conducted during the prior 30 days, for which I could obtain internals and re-weight using an average of the 2004 and 2008 partisan mix (except that I use Gallup's and Rasmussen's numbers without adjustment), is 2.3 points, with a MOE of 0.4 points. In the 42 states in which voter register by party, the trend since 2008 has been to reduce the advantage enjoyed by the Democrats. This trend has been especially strong in the battleground states. Since 2008, in both the Gallup and Rasmussen polls, the trend in party affiliation has been strongly in favor of the Republicans. Both the Exit Poll of 2008 and the post-election Current Population Survey (CPS) of the U.S. Bureau of Labor Statistics showed surges in black voter participation and young adult voter participation (with both groups heavily favoring Obama). It is unlikely that young adult voter participation will be like it was in 2008 (or that young adults will heavily favor Obama). The Gallup Poll and various election officials indicate a surge in early voting by Republicans relative to 2004. For that matter, the Republican GOTV effort has focused on marginal Romney voters, not merely banking the Romney vote. Republicans enter the last week of the election in a much stronger cash position that the Democrats.As the challenger in this election, we can suspect that the few remaining undecideds will break to Romney. Dr. Thies is a professor of statistics and econo-metrics at the Univ. of Shenandoah in Virginia.
NH +2.2 (10 surveys, MOE 1.3)NV +2.4 (19 surveys, MOE 0.9) OH +2.1 (37 surveys, MOE 0.7) VA +1.9 (26 surveys, MOE 0.8) TOSS-UP IA -0.1 (10 surveys, MOE 1.3) WI -0.7 (17 surveys, MOE 1.0) LEANS OBAMA MN -1.4 (8 surveys, MOE 1.4) PA -1.0 (16 surveys, MOE 1.0) PROBABLE OBAMA MI -3.8 (8 surveys, MOE 1.4) OR -3.1 (4 surveys, MOE 2.0)
Friday, November 2, 2012
I'm calling it for Romney
Electoral College vote 285 to 237 Clifford F. Thies If Karl Rove can stick his neck out and call the race at this time, so can I. I'm calling it for Romney, 285 to 237 Electoral Votes, with Iowa (6 EVs) and Wisconsin (10 EVs) pure toss-ups. My call is very much the same as Rove's, except that I award Nevada to Romney and leave Iowa and Wisconsin as pure toss-ups, which three states he awards to Obama. With regard to all the close states, here is my specific reasoning, based on state polls conducted during the prior thirty days for which I could obtain internals and re-weight using an average of the 2004 and 2008 partisan mix (except that I use Rasmussen's numbers without adjustment):