John F. Kerry gained ground on George W. Bush in Ohio after Election Day.  The official results, posted on December 2, 2004, include provisional ballots and some late-arriving absentee ballots.  The point spread between Bush and Kerry has been reduced by 0.30%, but this only begins to tell the story.

All the following tables show the official results on the first line, the unofficial results on the second line, and the difference on the third line.  The number of newly counted votes as a percentage of the total is posted beneath the name of the county.  State wide, this percentage was 2.62%.

COMPARISON OF RESULTS: STATE WIDE                

               Bush    % Bush     Kerry   % Kerry
Total      2,858,727   50.82%  2,739,952   48.70%
2.62%      2,792,244   50.97%  2,659,664   48.55%
              66,483   45.10%     80,288   54.46%

Note that I have adjusted these numbers to account for Gahanna, Ward 1, Precinct B, in Franklin County, where 638 ballots were cast, but Bush was awarded 4,258 votes.  Kerry received 260 votes.  There were 87 reported write-in votes, compared to zero in the rest of the ward, 13 in the rest of Gahanna.  These have now disappeared from the count.  Bush’s official total is now 365, and I have reduced his unofficial total by 3893, in order to make valid comparisons.

For less than 3% of the votes counted to offset by 0.30% the margin between the two candidates is significant.  State wide, Kerry’s percentage of the newly counted votes was 54.46%, much greater than his percentage in the officially certified statewide results.  The newly counted votes, whether absentee or provisional, were all paper ballots.  The sample size is statistically significant.  If the newly counted ballots do not represent a random sample of the electorate, it can only be because, on the one hand, absentee ballots tend to break in favor of Republicans, and on the other hand, provisional ballots disproportionately favor Democrats because of Republican challengers at the polling places.

There are 88 counties in Ohio.  The official count increased Bush’s percentage of the vote in only 16 counties, all of them small, the largest being Belmont County:

County        Bush    % Bush    Kerry   % Kerry    
Belmont      15,589   46.79%    17,574   52.74%
1.91%        15,275   46.74%    17,256   52.80%
                314   49.37%       318   50.00%

One of these 16 counties, Harrison County, is suspect because the vote totals actually decreased after Election Day, by an equal amount for both candidates: 

County        Bush    % Bush    Kerry   % Kerry   
Harrison      4,274   52.71%     3,780   46.61%
-1.09%        4,318   52.61%     3,824   46.59%
                -44   44.44%       -44   44.44%

Bush’s largest percentage gain was in Perry County, where the sample of newly counted votes is small, and the election results are suspect, as detailed in a letter dated December 2, 2004 from Congressman John Conyers and others to Ohio Secretary of State Kenneth Blackwell.  Irregularities included voter turnout of 124.4% in Precinct Reading S and 124.0% in Precinct West Lexington G, 3100 voters listed as having registered on the exact same day, and substantial numbers of voters with no signatures on file.

County        Bush    % Bush    Kerry   % Kerry   
Perry         7,855   51.72%     7,257   47.78%
0.90%         7,767   51.62%     7,205   47.88%
                 88   62.41%        52   36.88%

Another county where Bush’s percentage increased was Van Wert County, where the official turnout in Precinct Van Wert 4C was zero.  According to a letter I received on December 5, 2004 from the Chairman and Director of the Van Wert County Board of Elections, this was done under the direction of J. Kenneth Blackwell’s office.  The reason for these zero results “is due to the fact that precinct officials borrowed a voting booth from Precinct Van Wert D and the rotations, as required by law, for these opposed candidates in the two precincts are not the same.”  Thus, in the same election, Blackwell has certified voter turnout percentages at the precinct level ranging from a high of 124.4% to a low of zero.

County        Bush    % Bush    Kerry   % Kerry   
Van Wert     10,678   72.02%     4,094   27.61%
1.77%        10,484   71.99%     4,026   27.64%
                194   74.05%        68   25.95%

Other than Perry County, Bush’s largest percentage gain was in Ottawa County, where his share of the newly counted votes was 4.20% higher than the results reported on election night:

County        Bush    % Bush    Kerry   % Kerry   
Ottawa       12,075   51.92%    11,114   47.79%
1.96%        11,820   51.84%    10,915   47.87%
                255   56.04%       199   43.74%

On the other hand, the official count increased Kerry’s percentage of the vote in 66 counties (in 6 counties the percentages were unchanged).  Kerry’s share of the newly counted votes was more than 4% higher than his unofficial total in 36 counties, including the ten biggest counties in the state:

COMPARISON OF RESULTS: TEN BIGGEST COUNTIES

County         Bush    % Bush    Kerry   % Kerry   
Cuyahoga     221,606   32.89%   448,486   66.56%
3.15%        215,624   33.05%   433,262   66.41%
               5,892   27.72%    15,224   71.62%
Franklin     237,252   45.17%   285,800   54.35%
3.29%        230,293   45.29%   275,573   54.19%
               6,959   40.27%    10,227   59.18%
Hamilton     222,404   52.50%   199,499   47.09%
3.63%        215,639   52.82%   190,956   46.78%
               6,765   44.11%     8,543   55.71%
Montgomery   138,361   48.97%   142,977   50.60%
2.97%        134,716   49.14%   138,262   50.43%
               3,645   43.37%     4,715   56.10%
Summit       118,553   42.91%   156,578   56.67%
2.20%        116,184   42.99%   152,897   56.58%
               2,369   39.14%     3,681   60.82%
Lucas         87,106   39.56%   132,536   60.19%
2.44%         85,405   39.76%   128,874   60.00%
               1,701   31.62%     3,662   68.08%
Stark         92,211   48.93%    95,338   50.59%
2.88%         89,859   49.10%    92,295   50.43%
               2,352   43.24%     3,043   55.95%
Butler       109,866   65.87%    56,234   33.86%
3.12%        106,735   66.05%    54,185   33.53%
               3,131   60.14%     2,049   39.36%
Lorain        61,196   43.48%    78,965   56.11%
2.78%         59,751   43.67%    76,512   55.93%
               1,445   36.96%     2,453   62.74%
Mahoning      48,712   36.70%    83,069   62.58%
1.76%         47,968   36.79%    81,500   62.50%
                 744   31.96%     1,569   67.40%

These results are profoundly disturbing.  The newly counted votes, all with a certifiable paper trail, consistently show a much higher percentage for John Kerry than the results tabulated electronically on election night.  The difference is 8.93% in Hamilton, 8.08% in Lucas, 6.81% in Lorain, 5.83% in Butler, 5.67% in Montgomery, 5.52% in Stark, 5.21% in Cuyahoga, 4.99% in Franklin, 4.90% in Mahoning, and 4.24% in Summit.  These numbers indicate a systematic underreporting of the Kerry vote on election night.  Either the discrepancy must be entirely attributable to absentee ballots and provisional ballots coming disproportionately from Democratic precincts, or else the electronic vote counts were not correct.  Either way, the discrepancy is so pervasive that it cannot be easily dismissed.  These are the ten biggest counties in the State of Ohio.  Together they accounted for 61.30% of Kerry’s state wide vote.

The same pattern, a discrepancy of 4% or more between the election night results and the newly counted votes, shows up in 22 smaller counties, listed in order according to the size of the discrepancy: Clark, Marion, Muskingum, Morrow, Richland, Allen, Wood, Morgan, Jackson, Scioto, Holmes, Hancock, Defiance, Clinton, Portage, Shelby, Meigs, Champaign, Licking, Noble, Greene, and Wayne.

There are some other interesting patterns that emerge when comparing the official and unofficial results.  In two counties the new votes counted as a percentage of the total is extraordinarily high:

County         Bush    % Bush    Kerry   % Kerry   
Athens        10,847   36.10%    18,997   63.31%
8.59%          9,912   36.09%    17,369   63.24%
                 935   36.30%     1,628   63.20%
Guernsey       9,961   55.84%     7,768   43.54%
8.81%          9,095   55.93%     7,072   43.49%
                 866   54.84%       696   44.08%

Athens County, where 2504 provisional ballots were issued, is the home of the Main Campus of Ohio University, which suggests relentless challenging of the right of college students to vote.  In Guernsey County only 558 provisional ballots were issued, and I do not know why so many votes were counted late.

In Lake County, the number of newly counted votes split evenly, exactly, 1578 votes for Bush, and 1578 votes for Kerry.  This is plausible, given that the overall vote in Lake County was very close.

County         Bush    % Bush    Kerry   % Kerry   
Lake          62,193   51.05%    59,049   48.47%
2.60%         60,615   51.09%    57,471   48.44%
               1,578   49.76%     1,578   49.76%

I did ask to see the unofficial results for Lake County in order to analyze the provisional ballot count by comparing the precinct totals.  I was advised by the Board of Elections that the precinct totals are unchanged from the unofficial results.  Because of electronic voting, the provisional ballots and the absentee ballots are counted on an at-large county wide basis and lumped together in one line in the official election results.

Finally we come to Sandusky County.  There are 39,408 registered voters.  According to the unofficial results there were 32,433 ballots cast, a turnout of 82.30%, the highest percentage of any county in the state (Delaware County being the next highest at 79.16%).  In addition there were 760 provisional ballots issued, which could have brought the turnout to 84.23%.  There were 31,846 votes counted for president, which left 587 regular ballots uncounted.

Now, according to the official results, there were only 28,955 votes counted for president, a net loss of 2,891 votes.  More votes than this must have been disqualified, because some of the 760 provisional ballots must have been valid.  This amounts to a loss of at least 9.08% and as much as 11.20% of the vote.

County         Bush    % Bush    Kerry   % Kerry   
Sandusky     16,198   55.94%    12,653   43.70%
-9.08%       17,824   55.97%    13,909   43.68%
             -1,626   56.19%    -1,256   43.40%

The reason for this was reported by Jay Cohen of the Associated Press.  Barb Tuckerman, Director of the Sandusky County Board of Elections, stated that about 2600 ballots from nine precincts were counted twice when they were mistakenly placed alongside a pile of uncounted ballots.  The room where the ballots were being fed into optical-scan machines on election night was so crowded that ballots were being placed on the floor.  “It was totally hectic,” she said.  She further stated that no problems were found related to the machines that tally the paper ballots.

This explanation can only be scrutinized by comparing the unofficial results to the official results.  Once we have both sets of numbers, we intend to do so.