|The 2004 Election: Analyses, Summaries, Charts, and Spreadsheets|
2004 Ohio Presidential Election: Cuyahoga County Analysis
This page roughly compiles diverse writings about non-votes and undervoting
Many votes in Ohio were counted for the wrong candidate or ballot option, and a significant portion of the cross-voting ballots resulted in Kerry votes tallied as Bush votes. Cross-voting also impacted other races and analysis of past elections may reveal cross-voting.
E-voting and Non-Votes. With outcome-altering programming, a set percentage of one candidate's support may be discounted electronically. With other technologies, like punch card voting, undercounting voting is likeler to subtract from the candidates at their respective rates of support.
In NM 2004, the non-votes are e-voting, not op-scans. Only 334 non-votes of 18,731 were op-scanned. Even though analysis of the precinct candidate support levels, restoring normal non-voting and using precinct voting to reassign the excess non-voting (in effect giving precincts % counted parity) indicates Bush won, the e-voting equipment can be programmed to not count a percent of only one candidates votes. Under this presumpton, Kerry could have carried NM!
Spoilage can increase non-voting. Punching out a dozen ballots at once for candiate x spoils every ballot in the stack voting for any other candidate, but does not spoil any x voting.
Spoilage increases if ballots and equipment misalign. Punched out paper clogs machines, making correct ballot insertion problematic.
To keep elections fair, Ohio candidate names rotate to the top of the ballot list an equal portion of the time, a statuatory requirement. Election officials ensuring this equal rotation combined differing ballot orders at locations shared by several precincts, creating a greater problem than the one the law seeks to solve. In punch card voting Ohio counties I define "wrong-precinct voting" as ballots cast at a voting machine for one precinct then counted with a ballot order for a different precinct.
The Ohio ballot order was sequenced alphabetically and originally included Ralph Nader. Nader was converted to "disqualified" and the original alphabetic order was maintained. Table 1 shows the five ballot orders and the sequence used in 2004. The number of votes for d, "disqualified," was not reported by Cuyahoga County election officials. The official results indicated the number of recorded votes for the four candidates only and the number of ballots cast. Hence d, "disqualified," is an uncounted catch all for no selection, hanging chads, and cross-votes. Herein, the term "non-vote" refers to d, the "disparity" between the reported number of ballots cast and the number of votes counted in the race. The term "undervote" is a synonym. Retaining an empty slot in the order and not reporting voting for the option is unique. In the Cuyahoga cross-voting context, the non-vote has unique implications; not reporting their tally obfuscates evidence of voting irregularities.
Non-votes also include votes not counted (i.e. hanging chads) and no vote cast for president. In some states, the non-vote percentage is well below one percent. Ohio recorded near the highest rate in 2004. The Ohio electronic vote percentage is near normal, just 0.76%, compared to 1.90% for punch card votes. In punch card counties, a substantialy higher portion of the Ohio vote was not counted. Additionally, many cross-votes were counted opposite the voter's intention.
Table 12 presents the precinct average number of non-votes, precincts sorted by Kerry support.
Figure 2 illustrates the rate of non-votes for the 2\2 precincts sorted by the level of Kerry support. At 2\2 locations, when Kerry cross-votes switch to Bush in one precinct they switch to disqualified from the other precinct. Assuming random, equal cross-voting in both directions, the non-vote percentage is a useful and direct indicator of Kerry-Bush vote-switching. The non-vote rate is highest for precincts with a preponderance of Kerry support, that is, where cross-voting reduces the Kerry vote the most and vote-switching most heavily favors Bush.
It appears Kerry was robbed of votes!
Figure 3 shows the Cuyahoga precincts with greater than one standard deviation (z-score > 1) from the mean non-vote percentage, 1.80% in Cuyahoga County. Z-score is the number of standard deviations from the mean. These are the Cuyahoga precincts with over 3.6% non-votes, including those with only one ballot order. This sort represents 11.0% of the ballots cast and 27.7% of the non-votes.
Nearly all of these are precincts with over 75% Kerry support, most have over 90% Kerry support! The 11.0% of Cuyahoga precincts with over 3.6% non-votes have 27.7% of the non-votes and a mean of 10.25% Bush voting, so only one of ten uncounted votes are probable Bush votes. The precincts with greater than 1.5 stdev have a median of 94% Kerry support (Figure 4). With non-votes, what really matters is, "How many of who's votes are uncounted." Where votes go uncounted is more significant than how many. In Cuyahoga County, Kerry lost a very high, non-random proportion of the uncounted votes. This concentration of cross-voting/undercounting where it impacts results the most raises suspicion of irregularities and presents yet another mystery, "Why is undervoting concentrated in locations with the highest Kerry support?"
In Ohio, Bush's reported margin is greater than the non-vote total. If the uncounted vote in New Mexico, the nation's highest at 2.42%, was this one-sided (over 75% uncounted Kerry voters), Kerry won in NM. Officially, Bush won NM by 6,000 votes with 18,000 non-votes. Under random circumstances, when candidates lose votes proportional to the results, non-voting does not change the outcome. The Kerry to non-vote correlation for 33 NM counties is 0.454, the Cuyahoga County correlation for 1432 precincts is 0.423.
If the Ohio e-vote non-vote percentage is applied to Cuyahoga County precincts, and if the excess non-votes over that percentage are counted according to the reported precinct voting, the added votes would tally 4,833 Kerry and 1,284 Bush. The 11.0% of Cuyahoga precincts with over 3.6% non-votes have 27.7% of the non-votes. Kerry's support county wide is 65.66% while 79.01% of excess non-votes are untallied Kerry supporters. The Bush vote was reported to be 32.00% while only 20.99% of the untallied are Bush supporters. In NM, this proportion, 79% Kerry to 21% Bush, would give Kerry the victory. Simply put, "While Cuyahoga reported 2/3 Kerry voters, 4/5 of excess non-votes are Kerry supporters." Why is undercounting focused in Kerry areas? In Cuyahoga, part of the non-vote is cross-voting.
The big difference between Ohio and NM seems to be
I wrote, "If New Mexico had Nevada 's undervote rate and 3/4 of the remaining undervotes are uncounted Kerry votes, Kerry wins New Mexico ." This comparison can be made using, instead, the NM counties with normal undervotes. All the counties below the 1% line on the graph below average, at 0.45%, even lower than Nevada 's county average. Such a comparison also emphasizes the unfair NM machine bias (or whatever this was). I just studied this further, updating the NM spreadsheet in the process. Here is a new statistical summary of the vote counting disparities in NM, comparing the two very obvious NM groups:
In the low undervote group of 12 counties, Bush has 19% higher support, and all 12 of the normal undervote counties reported favoring Bush.
The Kerry support correlation to undervoting is equally explicit in Cuyahoga County. The % Kerry to % non-vote correlation for 33 NM counties is 0.454, the Cuyahoga County correlation for 1432 precincts is 0.423. Bush's vote has a -0.562 correlation to non-vote percentages for NM counties, -0.500% in Cuyahoga precincts. The big difference between Ohio and NM seems to be which minority is being robbed of votes.
If the Ohio e-vote counties' percentage of non-votes is applied to Cuyahoga County precincts, and if the excess non-votes over that percentage are counted according to the reported precinct voting, the added votes would tally 4,833 Kerry and 1,284 Bush. The 11.0% of Cuyahoga precincts with over 3.6% non-votes have 27.7% of the non-votes and a mean of 10.25% Bush voting. Kerry's support county wide is 65.66% while 79.01% of excess non-votes are undertallied Kerry supporters. The Bush vote was reported to be 32.00% while only 20.99% of the undertallied are Bush supporters. In NM, this proportion, 79% Kerry to 21% Bush, would give Kerry the victory. Simply put, "While Cuyahoga reported 2/3 Kerry voters, 4/5 of excess non-votes are probable Kerry supporters."
Why is undervoting focused in Kerry areas? In Cuyahoga, part of the non-vote is punch card cross-voting. In NM, punch cards cannot be blamed and the correlation to voting machines is far beyond credibly coincidence. The NM officials who failed to properly respond to a grossly unfair election have yet to justify their inaction.
RETURN TO: How Kerry Votes
Were Switched to Bush Votes
© 2008 All
Rights Reserved by James Q. Jacobs.