After the jump are my projections for Indiana extrapolated to a county-by-county level, along with some basic demographic information about each county. Since returns come in by county rather than Congressional District -- I figure it will be a little easier to follow the action this way. However, because my model was not designed to work at the county level, there may be an additional margin of error here. The results for the large counties should be more reliable than the results for the smaller ones.
Very important: These results are calibrated in order to match the current margin in the Real Clear Politics polling average: Clinton +5.0. For Obama to win Indiana, he'll need to be beating these numbers by about 5 percent across the board. For Clinton to win Indiana by double digits, she'd need to beat these numbers by about 5 percent.
Tuesday, May 6, 2008
Indiana Scorecard
Subscribe to:
Post Comments (Atom)
6 comments
I would hop Obama does better than +2 in Lake County. With Gary I think he probably should.
Regression Model using Pop_Black, Pop_65+, Pop_Bachelors, Pop_18-29, MedianIncome, Pop_Male and county-by-county results combined from Ohio and Pennsylvania.
Kerry 2004 totals were used to extrapolate final vote totals
Clinton - 516,200
Obama - 428,400
Clinton WIN by 9.30%
County Clinton Vote
Adams 65.1%
Allen 53.7%
Bartholomew 58.5%
Benton 65.8%
Blackford 68.5%
Boone 54.0%
Brown 62.2%
Carroll 64.4%
Cass 65.3%
Clark 59.4%
Clay 65.8%
Clinton 65.7%
Crawford 68.2%
Daviess 67.5%
Dearborn 59.4%
Decatur 64.1%
DeKalb 62.8%
Delaware 58.2%
Dubois 62.2%
Elkhart 57.9%
Fayette 67.1%
Floyd 58.0%
Fountain 67.4%
Franklin 63.3%
Fulton 66.2%
Gibson 63.4%
Grant 62.6%
Greene 67.6%
Hamilton 36.9%
Hancock 54.1%
Harrison 62.7%
Hendricks 51.4%
Henry 66.3%
Howard 58.0%
Huntington 64.4%
Jackson 64.5%
Jasper 63.0%
Jay 67.7%
Jefferson 63.6%
Jennings 65.0%
Johnson 56.3%
Knox 66.8%
Kosciusko 61.3%
LaGrange 63.4%
Lake 44.0%
LaPorte 58.8%
Lawrence 67.0%
Madison 61.3%
Marion 40.0%
Marshall 63.1%
Martin 67.6%
Miami 63.3%
Monroe 44.8%
Montgomery 64.1%
Morgan 61.1%
Newton 65.2%
Noble 63.6%
Ohio 64.9%
Orange 68.7%
Owen 67.0%
Parke 66.2%
Perry 66.6%
Pike 68.0%
Porter 55.1%
Posey 60.9%
Pulaski 66.9%
Putnam 62.8%
Randolph 67.9%
Ripley 64.4%
Rush 65.5%
Scott 67.5%
Shelby 62.6%
Spencer 63.6%
St. Joseph 52.3%
Starke 67.7%
Steuben 63.6%
Sullivan 65.9%
Switzerland 67.9%
Tippecanoe 44.7%
Tipton 62.4%
Union 66.2%
Vanderburgh 59.4%
Vermillion 67.0%
Vigo 58.9%
Wabash 65.4%
Warren 63.8%
Warrick 56.8%
Washington 66.7%
Wayne 64.3%
Wells 64.1%
White 66.5%
Whitley 62.7%
Hm. Steuben and Vigo are both reporting 67% or so and you nailed Vigo, Poblano Steuben...
These county by county projections are a fantastic way to keep track of how the candidates are doing vs. expectations (C+5%). Looks like your county projections are remarkably accurate with CNN reporting 36% precincts reporting. Obama seems to be doing a little better in his counties and Clinton seems to be significantly outperforming in south Indiana (Limbaugh effect?). All else seems fairly close.
If Hillary won Indian and Pennsylvania, those residents IQs have suddenly dropped! What are you people thinking. What makes you think Hillary is in it for YOU??? Can't you all see she hasn't accomplished anything being senator of NY? or as wife of a former president? Everything she has done in the state of NY has hurt New Yorker working class. What part of "Obama" and "Get over yourselves" (to PA residents) do you not understand???
Poblano,
I compared your model's margins to the final margins (rounded, a bit lazily). The results, at intervals of 3%, came out like this:
http://img182.imageshack.us/img182/4110/indianaregressea2.png
Clearly, the model seems to far, far understimate Clinton in the southeastern part of the state. I doubt this is a coincidence. Will you be adjusting it for Kentucky?
Most of the dark-green counties seem to have an explanation. Brown has an arts colony. Benton had remarkably low turnout; only 4% above Kerry, the second-lowest relative gain in the state. The rest are a little unclear. Montgomery seems to have a small college. White and especially the highly populated Elkhart are surprising though.
Great model, and I hope you continue to develop it, considering the new results in future analysis. I especially look forward to seeing it out West where cultural voting seems a little murkier.
Thanks as always for your excellent work.
Post a Comment