Constructing a baseline for Florida results presents much more difficult challenges than those posed by any state that has thus far gone to the polls.
In Iowa, in both 2008 and 2012, Mitt Romney was the less conservative establishment candidate who was up against a more conservative alternative whose base was found primarily among evangelicals (Huckabee in 2008, Santorum in 2012).
In New Hampshire in both 2008 and 2012, Romney was the more conservative, establishment candidate running against the more mavericky, left-leaning, and independent leaning alternative (McCain in 2008, Paul + Huntsman in 2012).
In South Carolina, Romney stepped into the shoes of McCain, finding his strongest support among moderates along the coast and in the Midlands, running against the more conservative alternative whose support was concentrated among evangelicals and very conservative voters (Huckabee in 2008, Gingrich in 2012).
Florida is altogether more difficult to evaluate.
Follow me over the jump for the analysis, or just click the below to go straight to the model
http://tinyurl.com/...
In 2008 Romney’s base of support was among the most conservative portions of the electorate while McCain’s base was among the more moderate segments (in 2008 Romney won very conservative voters (Romney/McCain/Huckabee/Giuliani) by a margin of 44/21/20/12 while losing moderates 21/43/8/21). –
http://www.cnn.com/...
This time around the relative bases of support are completely flipped with Romney strongest among moderates and weakest among the most conservative members of the party. In addition, while the South Florida Hispanic vote was strongly behind McCain and Giuliani in 2008, there are tentative signs that it is lining up behind Romney this time around (note Romney’s commanding lead among Cubans in the Univision poll and his support from the three Cuban Republicans representatives in South Florida) –
http://race42012.com/...
http://www.tampabay.com/...
This reasoning is backed by the headlining Florida 2010 Senate primary where the tea-partier Rick Scott narrowly beat establishment (relative moderate) Bill McCollum based off of Scott’s support in Romney’s strongest areas in 2008 (Jacksonville, SW Florida, the Panhandle) and in the face of opposition in Romney’s weakest areas (Miami-Dade and South Florida in general).
http://www.tampabay.com/...
Moreover both the 2008 and 2012 races featured another very conservative alternative whose base of support was primarily among evangelicals (Santorum is, in fact, is polling at about the same clip that Huckabee ultimately achieved in 2008). However, unlike in 2008, there is not moderate alternative akin to Giuliani and the ever enigmatic Ron Paul is set to take a much larger portion of the vote than he did in 2008.
The most logical step is therefore the most counter-intuitive, we should expect that Romney’s strongest areas in 2008 should be his weakest areas in 2012 (the Jacksonville area, Southwest Florida, parts of the Panhandle) and his weakest areas in 2008 should be his strongest areas in 2012 (Miami-Dade, and South Florida specifically, and likely to a lesser extent in the I-4 corridor).
With this in mind I’ve constructed a baseline for Florida primary below. It assumes that Paul gets 10% of the vote distributed roughly evenly across the state, Santorum matches Huckabee’s 2008 performance, Romney substantially underperforms the McCain+Giuliani vote from 2008 and Gingrich roughly matches, or marginally improves upon, Romney’s 2008 totals. Although these assumptions are very aggressive, they’re ultimately what’s required to get the state to a 50/50 between Gingrich and Romney. If this were to occur you would expect Gingrich to run up big margins in North Florida and SW Florida, lose badly in S. Florida (and Miami-Dade in particular) and to roughly split the I-4 corridor.
I should, point out that there are two areas where I think this model is somewhat imperfect. The first is that Santorum seems to have relatively weaker support than Huckabee did in 2008 among evangelicals and relatively stronger support among the rest of the electorate. The result of this is that the model probably understates Gingrich’s total in the 50/50 scenario in the Panhandle and overestimates it in the rest of the state.
The second is that I expect the model overstates Romney’s support in Miami-Dade (I expect Gingrich to lose the Cuban vote badly but to do better than the model would indicate) and that it understates Romney’s vote in the Jacksonville suburbs and in SW Florida since both areas have a primary electorate that is not only quite conservative but also quite high-income (a demographic Romney did well with in both 2008 and 2012).
Should Romney beat Gingrich by 5-10 points I would expect that Gingrich would hold on to much the Panhandle while drawing SW Florida and the Jacksonville area to a 50/50, while losing the I-4 corridor by a moderate amount and while getting smashed in S. Florida even worse than this model suggests.
Should Romney beat Gingrich by 15-20 points I would suspect Gingrich would be losing most everywhere except perhaps for Bay County (Panama City) and few rural Panhandle and Florida Heartland counties.
Without further discussion, please note a summary of key counties below (with apologies for the erratic spacing) and a google docs link to the full calculations…
http://tinyurl.com/...
Rom Gin San Paul
Alachua 30 34 21 15
Bay 28 42 20 10
Brevard 37 39 14 10
Broward 48 31 11 10
Clay 25 49 16 10
Collier 32 51 7 10
Duval 26 48 16 10
Escambia 30 37 23 10
Hillsborough 40 37 13 10
Lee 34 47 9 10
Leon 36 32 17 15
Manatee 38 40 12 10
Miami-Dade 63 11 6 10
Nassau 26 43 21 10
Okaloosa 33 38 19 10
Orange 34 39 17 10
Osceola 33 38 19 10
Palm Beach 44 36 9 10
Pinellas 41 38 11 10
Santa Rosa 31 35 24 10
Sarasota 41 40 9 10
Volusia 35 41 14 10
Walton 31 37 22 10