Mitt Romney 46.42% 775,014
Newt Gingrich 31.93% 533,117
Rick Santorum 13.34% 222,799
Ron Paul 7.01% 117,104
Others 1.29% 21,613
My projection from last night:
Romney: 44.02%
Gingrich: 32.15%
Santorum: 13.19%
Paul: 9.64%
It looks like I'm going to underestimate Mitt's resul by 2% and overestimate Ron Paul by about 2%. All in all pretty close.
Since individual polls do not allocate undecided votes while I do, it is more appropriate to look at margins than the % totals overall.
Looking at the latest polls by each pollster that fell within the last week, a fair comparison can be made.
Overall by margin difference:
Suffolk University: Romney +20
We Ask America: Romney +19
Rasmussen: Romney +16
Survey USA: Romney +15
NBC News/Marist: Romney +15
Quinnipiac: Romney +14
Mason-Dixon: Romney +11
American Research Group: Romney +11
War Room Logistics: Romney +10
Voter Survey Service: Romney +9
Ipsos/Reuters: Romney +8
Public Policy Polling: Romney +7
Monmouth: Romney +7
Insider Advantage: Romney +5
CNN/Time: Romney +2
Florida Chamber of Commerce: Tie
Looking at this, 6 pollsters overestimated Romney's win and 10 underestimated it, although considering allocation of undecideds probably would have pushed Mason-Dixon & ARG to the right margin.
Another interesting point is that 12 of the 16 polls overestimated Ron Paul, this is especially damning when considering that undecideds typically range from 6-10% in a given poll, therefore typically pollsters should underestimate an individual candidate. A clear example is that only 1 poll had Mitt Romney over 46% (his actual result), while the other 15 pollsters showed him no higher than 44%.
No comments:
Post a Comment