The 2004 Election:  Analyses, Summaries, Charts, and Spreadsheets

Ohio 2004 Presidential Elections: Results, Summary, Charts and Spreadsheets
© James Q. Jacobs. All Rights Reserved.

Scroll Down for the Charts and Results Summary

Return to: How Kerry Votes were Switched to Bush Votes

Cuyahoga County Precinct Level Analysis

NEW Access Database: Cuyahoga Results with Probability Sorts
cuyahoga_precincts.mdb — All Cuyahoga Precincts with Probability Sorting - 912 Kb

NEW July 2006: Updated Spreadsheets with Probability Sorting:
Cuyahoga County Official Results, with Location Subsets
Six Excel spreadsheets displaying results of statistical analysis.

NEW Jan. 22, 2007: Extended Analysis of Probability Subsets
Continued study following on cuyahoga_probability.xls above.


K-d represents a wrong-precinct vote changing from Kerry to disqualified.
The trend evidences the cross-voting. The higher the probability a Kerry wrong-precinct vote changes to disqualified, the higher the disqualified percentage.


Where is the cross-voting concentrated? Under random circumstances it would not be concentrated where the most Kerry votes are switched. This pattern is very suspicious!!!

NEW 2008: cuyahoga_probability_data.xls - K-B Probability Subsets - 0.7 Mb cuyahoga_probability_data.xls - Probability Subsets - 0.7 Mb
cuyahoga_3_precincts_subsets.xls - Probability Subsets - 0.28 Mb
cuyahoga_4_precincts_subsets.xls - Probability Subsets - 0.24 Mb

Data before adding Probability Sorting

The spreadsheets are part of my Cuyahoga County analysis and the data organization is explained therein. Read the article to understand the symbols used in the spreadsheets and charts.

The updated spreadsheets represent a new stage in the study; I restarted the analysis with probability sorting added. Missing data was found and inserted, and slight errors were corrected. Cross-vote probabilities for precincts and for locations were added, subset sorts were added, precincts and locations were resorted, probability sorts were analyzed and compared, and statistical summaries were added. Some significant statistical anomalies are highlighted. Charts of new results are incorporated into these pages.


The probability sorting reveals where cross-voting precincts were combined.
Precincts with the highest probability of Kerry cross-votes switching to Bush
are in areas with the highest Kerry support. Was this intentional manipulation?

The workbooks are ready for use in statistical programs and projects. I consider them reliable and they supersede my previous Cuyahoga County results files. Let me know if you detect any problem whatsoever. Joe Knapp did the initial heavy lifting, and gets loud applause for building, from the official County election text files, the original spreadsheet I used. I manipulated the data many times, and any errors are mine. Amidst the millions of numbers and digits, I hope there are no errors. I need to know if you find anything, no matter how slight.

Additional subset probability sorts by minor candidate probabilities have been added. This feature in cuyahoga_2.xls produced significant findings regarding intended third-party voting, and can serve as a model for other researh questions. In particular, note the changes in percentage of votes to minor party candidates when there is a 1.0 probability that switched Kerry votes will be counted for either minor candidate. In the worksheet "2_2_Locations" in cuyahoga_2.xls, compare the 1.0 probability percentages with the 0.0 probability precincts. At these 2 ballot order and 2 precinct locations, Peroutka received 1.932% of the vote in precincts where switched Kerry votes count for Peroutka, compared with 0.142% in precincts where switched Kerry votes count as Bush votes. In Jan. 2007, I uploaded an extended analysis spreadsheet, cuyahoga_precinct_subsets.xls with the same comparison for other subset, plus numerous charts.


Vote percentages are five to near 10 times higher for Badnarik and Peroutka in the
precincts with a 1.0 probability that Kerry cross-votes count for those candidates.
This indisputable, descriptive statistical evidence quantifies the cross-voting problem.

Again, to understand the spreadsheet notation, read "How Kerry Votes were Switched to Bush Votes." To view some of the latest statistical results, check the highlighting in the spreadsheet summaries. One reason the spreadsheet are released is so others can study the data without having to also do the tedious and time-consuming probability sorts. They are also a data foundation from which many questions, in addition to those herein, can be addressed.

Scroll down for tables, charts, and graphs.
Back to the Main Spreadsheets Page

The following table and chart display the state results sorted by county voting methods.
Some of my state-wide analysis preceded release of the official results.
Official Results are in the spreadsheets.

 2004 Ohio Election Results Summary
 
State
E-Vote
Op-Scan
Punch Card

Population

11,435,798
1,799,299
1,382,208
8,254,291

# Voters

7,979,639
1,344,131
925,806
5,709,702

# Votes

5,574,476
870,237
670,058
4,034,181

% Votes

 
15.61
12.02
72.37

% Cast

69.86
64.74
72.38
70.65

# Provisional

155,428
21,590
17,078
116,760

% Provisional

2.79
2.48
2.55
2.89

Kerry

2,659,664
451,975
293,998
1,913,691

Bush

2,796,147
406,697
363,895
2,025,555

Badnarik

14,331
2,967
1,177
10,187

Peroutka

11,614
1,986
990
8,638

Non-Votes

92,720
6,612
9,998
76,110

% Kerry

47.711
51.937
43.877
47.437

% Bush

50.160
46.734
54.308
50.210

% Badnarik

0.257
0.341
0.176
0.253

% Peroutka

0.208
0.228
0.148
0.214

% Non-Votes

1.663
0.760
1.492
1.887

Counties

88
7
13
68

Mean Bush Shift

0.903
0.615
0.536
1.003

How Kerry Votes were Switched to Bush Votes!
Several Cuyahoga county precincts (examples above) reported very unusual election results,
evidence of cross-voting, the casting of votes using equipment with the wrong ballot order.
As a result, many votes were counted for the wrong candidate or not counted.   Read the article.


The graph above depicts the statewide correlations of county non-vote percentages to six
other county variables, including candidate vote percentages. Additionally, the 88 Ohio
counties are subdivided into two groups, 44 above and 44 below the median county
non-vote percentage. In the counties above the mean, Kerry votes go down
in relation to the increase in non-votes, with a -0.194 correlation.
Correlations with third-party candidates indicate cross-voting.


This graph illustrates how the percentage of Bush vote change from 2000
to 2004 has the strongest correlation with Bush votes in punch card counties.
How much of the punch card correlation is attributable to vote-switching?

READ: The 2004 Ohio Presidential Election: Cuyahoga County Analysis
How Kerry Votes were Switched to Bush Votes! 


The chart above presents a comparison of the mean and median percentage of Kerry votes
for precincts with greater than and less than 1.5 standard deviations in the percentage
of non-votes. The precincts with the greater than 1.5 standard deviations in non-votes
are precincts with overwhelming Kerry support. This statistic infers that the greatest
amount of cross-voting occurred in precincts with the highest Kerry support.


This chart presents the differences in county non-vote percentages statewide and separated by
type of voting. The difference between e-voting and optical scan voting is notable. In
e-voting counties, the non-vote percentage is very consistent. In optical scan
voting counties, the standard deviation in non-vote percentages is very high.

Back to the Main Spreadsheets Page


The voting summary above presents questions about the different voting machine
types.  Why are the non-vote percentages so high in Op-Scan and Punch voting
counties?  Why do third-party candidates have highest support in E-Vote counties?

READ: The 2004 Ohio Presidential Election: Cuyahoga County Analysis
How Kerry Votes were Switched to Bush Votes!

© 2015. All Rights Reserved.    jqjacobs.net    |    Contact