OHIO 2008 – County by County - Part I
„forget the inverted C theory!“
Here is a county for county analysis of the Presidential Election 2008 for the state of OHIO.
Before OHIO, I just published a county by county analysis for the neighboring state of INDIANA. It is in three parts at Google Docs: Part I, Part II and Part III. This makes comparing the two states easier.
President Obama, who won the Buckeye state with a +4.58% margin, was the first democratic candidate to win this state since 1996. OHIO is currently the number one „bellwether state“ in the Union, having gone for the winning candidate every cycle since 1960, or, excluding the years 1944 and 1960, every cycle since 1896! Before 2008, MISSOURI was the number one bellwether state in the Union.
First, some important background information on OHIO:
Since the first appeared on the national ballot in 1856, there have been 39 election cycles, of which the democratic party has won 11 cycles (28.21%), and the GOP has won 28 cycles (71.79%). From 1856-1908, a span of 54 years, Ohio went exclusively for the GOP. Woodrow Wilson was the first democratic candidate to win Ohio. Every democratic candidate who has won Ohio has won the election. Only two democratic candidates in the 20th century captured the White House without Ohio: FDR (1944) and John F. Kennedy (1960). The conventional wisdom has been that a republican has never won the White House without Ohio (and this is statistically correct). For this reason, the buckeye state became the central focus of George W. Bush, Jr.'s 2004 re-election campaign and ended up being a fairly close race on election night.
An Obama win in Ohio in 2008 was anything but guaranteed. And as the democratic party learned from both 2000 and 2004, putting all your eggs in one basket is not such a great idea. However, the Obama campaign, through strategy and intensive field work, created a situation where it would have been possible for Obama to comfortably win the White House without Ohio: IOWA, NEW MEXICO, COLORADO, NEVADA and VIRGINIA were looking more and more like solid „locks“ for the democratic candidate as election day neared, and indeed those five states went for Obama with higher percentages and margins than Ohio itself. So, instead of being the central focus of the 2008 GE, Ohio rather swam with the general electoral stream.
Here 11 democratic winning cycles in OHIO:
|
Year |
Candidate |
D % |
R % |
I/O % |
Margin |
Part. Shift. |
|
1912 |
Wilson |
40.96% |
26.82% |
32.22% |
+14.14% |
+20.34% |
|
1916 |
Wilson |
51.86% |
44.18% |
3.96% |
+7.67% |
-6.47% |
|
1932 |
Roosevelt, F. |
49.88% |
47.03% |
3.09% |
+2.85% |
+33.28% |
|
1936 |
Roosevelt, F. |
57.99% |
37.44% |
4.57% |
+20.56% |
+17.71% |
|
1940 |
Roosevelt, F. |
52.20% |
47.80% |
0.00% |
+4.41% |
-13.30% |
|
1948 |
Truman |
49.48% |
49.24% |
1.28% |
+0.24% |
+0.61% |
|
1964 |
Johnson |
62.94% |
37.06% |
0.00% |
+25.89% |
+32.46% |
|
1976 |
Carter |
48.92% |
48.65% |
2.43% |
+0.27% |
+21.83% |
|
1992 |
Clinton |
40.18% |
38.35% |
21.47% |
+1.83% |
+12.68% |
|
1996 |
Clinton |
47.38% |
41.02% |
11.61% |
+6.36% |
+4.53% |
|
2008 |
Obama |
51.39% |
46.80% |
1.81% |
+4.58% |
+6.69% |
Here the same table, arranged by winning percent, descending:
|
Year |
Candidate |
D % |
R % |
I/O % |
Margin |
Part. Shift. |
|
1964 |
Johnson |
62.94% |
37.06% |
0.00% |
+25.89% |
+32.46% |
|
1936 |
Roosevelt, F. |
57.99% |
37.44% |
4.57% |
+20.56% |
+17.71% |
|
1940 |
Roosevelt, F. |
52.20% |
47.80% |
0.00% |
+4.41% |
-13.30% |
|
1916 |
Wilson |
51.86% |
44.18% |
3.96% |
+7.67% |
-6.47% |
|
2008 |
Obama |
51.39% |
46.80% |
1.81% |
+4.58% |
+6.69% |
|
1932 |
Roosevelt, F. |
49.88% |
47.03% |
3.09% |
+2.85% |
+33.28% |
|
1948 |
Truman |
49.48% |
49.24% |
1.28% |
+0.24% |
+0.61% |
|
1996 |
Clinton |
47.38% |
41.02% |
11.61% |
+6.36% |
+4.53% |
|
1976 |
Carter |
48.92% |
48.65% |
2.43% |
+0.27% |
+21.83% |
|
1992 |
Clinton |
40.18% |
38.35% |
21.47% |
+1.83% |
+12.68% |
|
1912 |
Wilson |
40.96% |
26.82% |
32.22% |
+14.14% |
+20.34% |
And the same table, arranged by winning margin, descending:
|
Year |
Candidate |
D % |
R % |
I/O % |
Margin |
Part. Shift. |
|
1964 |
Johnson |
62.94% |
37.06% |
0.00% |
+25.89% |
+32.46% |
|
1936 |
Roosevelt, F. |
57.99% |
37.44% |
4.57% |
+20.56% |
+17.71% |
|
1912 |
Wilson |
40.96% |
26.82% |
32.22% |
+14.14% |
+20.34% |
|
1916 |
Wilson |
51.86% |
44.18% |
3.96% |
+7.67% |
-6.47% |
|
1996 |
Clinton |
47.38% |
41.02% |
11.61% |
+6.36% |
+4.53% |
|
2008 |
Obama |
51.39% |
46.80% |
1.81% |
+4.58% |
+6.69% |
|
1940 |
Roosevelt, F. |
52.20% |
47.80% |
0.00% |
+4.41% |
-13.30% |
|
1932 |
Roosevelt, F. |
49.88% |
47.03% |
3.09% |
+2.85% |
+33.28% |
|
1992 |
Clinton |
40.18% |
38.35% |
21.47% |
+1.83% |
+12.68% |
|
1976 |
Carter |
48.92% |
48.65% |
2.43% |
+0.27% |
+21.83% |
|
1948 |
Truman |
49.48% |
49.24% |
1.28% |
+0.24% |
+0.61% |
Again, the same table, this time arranged by partisan shift, descending:
|
Year |
Candidate |
D % |
R % |
I/O % |
Margin |
Part. Shift. |
|
1932 |
Roosevelt, F. |
49.88% |
47.03% |
3.09% |
+2.85% |
+33.28% |
|
1964 |
Johnson |
62.94% |
37.06% |
0.00% |
+25.89% |
+32.46% |
|
1976 |
Carter |
48.92% |
48.65% |
2.43% |
+0.27% |
+21.83% |
|
1912 |
Wilson |
40.96% |
26.82% |
32.22% |
+14.14% |
+20.34% |
|
1936 |
Roosevelt, F. |
57.99% |
37.44% |
4.57% |
+20.56% |
+17.71% |
|
1992 |
Clinton |
40.18% |
38.35% |
21.47% |
+1.83% |
+12.68% |
|
2008 |
Obama |
51.39% |
46.80% |
1.81% |
+4.58% |
+6.69% |
|
1996 |
Clinton |
47.38% |
41.02% |
11.61% |
+6.36% |
+4.53% |
|
1948 |
Truman |
49.48% |
49.24% |
1.28% |
+0.24% |
+0.61% |
|
1916 |
Wilson |
51.86% |
44.18% |
3.96% |
+7.67% |
-6.47% |
|
1940 |
Roosevelt, F. |
52.20% |
47.80% |
0.00% |
+4.41% |
-13.30% |
And finally, the table to measure the partisan shift for the state over the national partisan shift for that year (this is often called „PARTISAN VALUE“):
|
Year |
Candidate |
Margin |
Part. Shift. |
Natl. Part Shift |
Diff. |
|
1936 |
Roosevelt, F. |
+20.56% |
+17.71% |
+6.49% |
+11.22% |
|
1964 |
Johnson |
+25.89% |
+32.46% |
+22.42% |
+10.04% |
|
1916 |
Wilson |
+7.67% |
-6.47% |
-11.32% |
+4.85% |
|
1948 |
Truman |
+0.24% |
+0.61% |
-3.01% |
+3.62% |
|
1996 |
Clinton |
+6.36% |
+4.53% |
+2.96% |
+1.57% |
|
1940 |
Roosevelt, F. |
+4.41% |
-13.30% |
-14.30% |
+1.00% |
|
1992 |
Clinton |
+1.83% |
+12.68% |
+13.29% |
-0.61% |
|
1932 |
Roosevelt, F. |
+2.85% |
+33.28% |
+35.18% |
-1.90% |
|
1912 |
Wilson |
+14.14% |
+20.34% |
+22.97% |
-2.63% |
|
2008 |
Obama |
+4.58% |
+6.69% |
+9.73% |
-3.04% |
|
1976 |
Carter |
+0.27% |
+21.83% |
+25.21% |
-3.38% |
Note that in 6 of 11 cycles, the difference is a postive difference. The other 5 cycles show a negative difference, a fairly even split.
So, based on these tables, here are undeniable facts about Obama's win in OHIO in 2008:
1.) Obama won OHIO with the largest raw vote total for a candidate of any party (2,940,044 votes, 80,267 votes more than Bush's record-breaking raw vote from 2004).
2.) Obama won OHIO with the 5th largest winning percent, the 6th largest winning margin and the 7th largest OHIO state partisan shift in US history for a democratic candidate. Of the 11 democratic winning cycles in OHIO, this puts Obama roughly in the middle of the standings.
§.) However, when we measure the difference between the statewide partisan shift over 2004 and the national partisan shift over the same („PARTISAN VALUE“), then Obama comes out near the bottom of the list, in 10th place. This means that the democratic tide in OHIO was less than in the nation as a whole, by -3.04%. And if you look at those negative partisan values, they are all for democratic candidates in their first term. The value for Clinton and Roosevelt changed to a positive value in their 2nd and 2nd / 3rd terms, respectively.
To put this in perspective, we should note that the last five election cycles (1992, 1996, 2000, 2004 and 2008) have all witnessed lean single digit wins for one side or the other in OHIO. In fact, if you go back to 1960, 8 of 13 election cycles in OHIO have been lean single digit wins for one side or the other. The margin average for the last five cycles is: +3.68%. For the last 13 cycles, it is: +8.86%. For this reason, we cannot expect massive partisan shifts in this state, as we flip from a lean win for one side to a lean win for the other. What we can say is that OBAMA won OHIO with more than double the margin with which Bush won OHIO in 2004. He also won with a larger margin than Bush's winning margin from 2000. Of the last five lean cycles, Obama's win in OHIO is the second largest, after Bill Clinton from 1996.
It is also interesting to note that the media networks, in spite of the leanness of the race, were all able to comfortably call Ohio for Obama at about 9:30 PM on election night, roughly 2 hours after the polls closed.
I took the county-wide data for the entire state for both 2008 and 2004 and plugged it into an excel document, which you can find here. To read the excel data, you may need to get a gmail account. Subsidiary data for all 88 counties in the state of OHIO which you can find here.
Of those 88 counties, Obama won 22 counties, or 25% (16 DEM retentions, 6 pick-ups), McCain won the other 66 (all GOP retentions.) In the excel document, everything is color-coded: red= GOP, blue = DEM, green= OTH. In the case of DEM pick-ups, those counties are also highlighted in light-blue.
And now comes the good stuff. Here is a quick table of the number of counties, by party, where there was a percentage loss or gain, a numeric (raw vote) loss or gain, and a margin loss or gain (margin loss or gain will not be applied to the „other“ vote“):
|
Party |
% loss |
% gain |
Num. loss |
Num. Gain |
Margin loss |
Margin gain |
|
GOP |
83 |
5 |
82 |
6 |
77 |
11 |
|
DEM |
16 |
72 |
17 |
71 |
11 |
77 |
|
OTHER |
0 |
88 |
0 |
88 |
-- |
-- |
Obama lost in %(over Kerry from 2004) in the following counties: Belmont, Clark, Columbiana, Gallia, Jackson, Jefferson, Lawrence, Mahoning, Monroe, Noble, Perry, Preble, Scioto, Trumbull, Vinton and Washington (16).
Obama lost raw votes (over Kerry from 2004) in the following counties: Adams, Belmont, Clark, Columbiana, Gallia, Guernsey, Harrison, Jackson, Jefferson, Lawrence, Mahoning, Monroe, Noble, Preble, Trumbull, Vinton, Washington (17). Note: It is possible to lose raw votes over the last cycle and still improve your percentage, depending on how badly your opponent did or if a third party spikes at the cost of your opponent :)
McCain lost in the raw vote over 2004 in most all counties. Here the counties where he actually improved in the raw vote over Bush from 2004: Darke, Delaware, Greene, Jefferson, Pickaway, Warren (6). It is possible to do better in the raw vote than your predecessor and still suffer a % loss, depending on how well your opponent does. Example: Warren County.
Likewise, here the counties where McCain improved his % over 2004: Belmont, Columbiana, Jefferson, Lawrence, Scioto (5).
This means that the GOP lost ground in 77, 82 or 83 counties, depending on how one wishes to measure it's loss. Since margins are used to measure shifts, the number 77 is the most useful number of the three.
For both parties, some of these losses or gains are statistically insignificant. For instance, in the case of the 7 counties where the margin shift was 1% or less. In Washington, Mahoning, Perry, Vinton, Jackson, Clarke and Preble counties, it is as if those counties froze in time while the other counties moved forward.
The 3rd and 4th party vote spiked mostly in very republican trending counties: Caroll, Coshochton, Gallia, Guernsey, Hardin, Harrison, Meigs, Marion, Noble, Paulding, Perry and Vinton counties. But it also spiked in a couple of democratic counties: Mahoning, Monroe. There can be no doubt that this spike was a form of protest vote, mostly against McCain, but also (in the Appalachian counties) also somewhat against Obama.
The numerical analysis shows that the growth rate over 2004 is mostly flat in OHIO, with occasional large spikes here. Most shifts by percent or margin are relatively gentle shifts, not the massive 20% or more shifts we saw in most of INDIANA. Gentle, but enough to put Obama over the top.