UPDATED: Spreadsheets for Cuyahoga County 2004 Presidential Tally

[L. Coyote]

The 2004 Ohio Presidential Election: Cuyahoga County Analysis
How Kerry Votes Were Switched to Bush Votes

http://jqjacobs.net/politics/ohio.html

The article and spreadsheets are being updated with more probability sorting. See:

http://jqjacobs.net/politics/xls/cuyahoga_2.xls (1+ MB).

for how new probability sorts define the actual vote of the third-party candidates.

[sinkingfeeling]

1. Thanks.

[Bill Bored]

2. Wow, a blast from the past, L. Coyote! There are some "newbies"

that might really be interested in this. They know who they are.

I hope you guys can talk about it because a lot of us have moved on to other work, but this is one that's always been intriguing.

[L. Coyote]

3. "....this is one that's always been intriguing."

and laborious.

However, the suspicious pattern--of which precincts had cross-voting and vote-switching--will not allow this to go away. Now, the new probability sorts just reinforce previous suspicions and conclusions: "stinks of fixed election"

The same probability sort now added to the spreadfsheets could be applied to the down-ticket races. All Ohio precincts with mutiple ballot orders should get this probability sort analysis. It will pinpoint vote-switching. But again, it is laborious.

RFK did not mention vote-switching, yet it reinforces the "stolen" premise better than the ephemeral voter who did not make it to the ballot box. Especially considering that SWITCHED VOTES COUNT TWICE, plus one for Bush AND minus one for Kerry.

[MissWaverly]

4. will there be any accountablility for this

is it more of Aw shucks, we'll do better next time. I know that my machine was rigged to
vote Republican in 2004, I have no confidence in either Bush or the GOP. Is there any hope
of correcting this.

[L. Coyote]

5. RE: "Is there any hope of correcting this?"

Only if someone actually "gets it." I have yet to see discussion of how the pattern in where cross-precinct voting did and did not occur was "created." Who designed the Cuyahoga election in such a skewed manner? There are so many irregularities, each one is kinda lost in the mix. But this one shows how part of the fix was in. The FIX iself was the cumulative impact of all manner of strategies, from scrubbing the roles, to creating long lines, etc. etc. The TOTAL impact is the fix. This one left a real trail of evidence, some did not.

The complexity of the myriad tactics and of the election organization itself obfuscates the irregularities. And it makes the work of revealing what happened very laborious and difficult. Probability sorting is a step forward. Separating the precincts where no cross-vote to candidate x can have occurred shows the true level of that candidates support, and the degree of cross-voting. Where the cross-voting happened reveals the intentionality.

What is the statute of limitations on fixing elections?

[Ms. Toad]

6. To refresh everyone's memory...

Cuyahoga County used the Florida hanging chad ballots. The ballots are not marked with any precinct designation (either machine or human readable). Counting the votes accurately on these ballots can only happen if the ballots are both punched and read using the same template.

Cuyahoga County (like most counties in Ohio) houses multiple precincts in a single location, each of which is required to have a different ballot rotation.

When the lines were long in some precincts, some voters were told by precinct workers to use voting booths that were assigned to a different precinct. They took their unmarked ballot card, punched it using the wrong template, walked their cards back to their own precinct ballot and inserted their unmarked, unsortable ballots into a box which was then counted against a different template. It's as if you took exam version A and accidentally marked your answer sheet which doesn't have the questions on it as exam version B (except worse, because there's no "A" or "B" mark on these ballots).

For example (assuming only two candidates): Precinct A lists the presidential candidates in the order : Kerry Bush; Precinct B lists the presidential candidates in the order: Bush Kerry. I am told by the precinct worker to take my ballot for precinct A, to a booth set up for precinct B. I punch the second hole (because Kerry is second on the precinct B list). I put it in my ballot box (precinct A) and go on my way happy to have voted for Kerry. When the machine set to read precinct A ballots reads a punch in the second hole it records my vote as a vote for Bush (since in Precinct A, Bush corresponds to the second hole).

Because the cards are unmarked, once they are intermingled there is no way to sort them back out to determine what template SHOULD have been used to count them.

This problem was discovered both by direct reporting (witnesses who saw precinct workers TELLING voters to use the wrong booth), and by unexpectedly high performance by third party candidates (voters punched the hole that lined up with Bush or Kerry - unfortunately, that hole on the counting template belonged to a third party candidate).

Unfair. It certainly skewed the results in Cuyahoga County and anywhere else this gross stupidity reigned (some pollworkers insisted that voters switch lines, even after being told by some of the voters that they were sure that it would mess up the count). I haven't reviewed the stats to see how big a problem it was (to see if the probable skew in Cuyahoga County would have been enough to flip the state), and I don't know that the analysis has been done in similar locations to see if it was more than an isolated problem. If either of the above (big numbers or beyond an isolated problem) applies, it ought to be grounds for a "do over," or some meaningful remedy.

Characterizing this particular problem as election fixing, however, is not particularly helpful, since the rotation is fixed by law (not designed by anyone who might want to "fix" the election), and election fixer would have needed to be engineer long lines in front of one particular booth in the precinct and short lines in front of another particular booth in the same precinct, in order to ensure that the votes rolled in the correct direction. All reports I heard indicated pollworkers were randomly directing (or voters were moving on their own) from whichever booth had the longest line to whichever had the shortest line.

When I have the chance I'll plow through the stats - this particular problem intrigued (and angered) me since I heard the first reports. I'd love to see how pervasive it was.

[L. Coyote]

7. RE: "Characterizing this particular problem as election fixing ..."

Have you read the entire article yet? You wrote: "Characterizing this particular problem as election fixing, however, is not particularly helpful, since the rotation is fixed by law..."

True, to a degree. The assignment of ballot orders followed a rotation, 1, 2 3, 4, 5, 1, 2, 3, etc., and was assigned to precincts in the precinct number order. This does not preclude many possible forms of manipulation, including how many precinct were at locations with more than one ballot order (over 7/8ths), which precincts were combined at those locations (Kerry and Bush in the same ballot locations or not), number of voters in each precinct at the location, assigning precinct numbers, combining precincts, etc., etc. This does not require a genius to set up (perhaps to unravel) and the complexity of the arrangement makes it practically self-obfuscating. Your statement is apologetic, at best, and an ill-informed conclusion.

Continuing, you write, "... election fixer would have needed to be engineer long lines in front of one particular booth in the precinct and short lines in front of another particular booth in the same precinct..."

What law prevented this? None. This is easy to engineer by assigning both numbers of voters and number of machines. Try analyzing that equation in relation to whether or not a Kerry vote would be switched to Bush at the location. All the data is there, ready to go.

Perhaps further analysis of the statistics will help you see the patterns. This was not a randomly generated situation. The probabilities were very skewed, favoring Bush, by the way the election was set up. Who did that? Not a law! Why is the skew in favor of Kerry votes being switched to Bush votes. Here is a more simple analysis. Compare what percentage of voters were subject to multiple ballot orders at their location in Cuyahoga or in all of Ohio with candidate support in those precincts!! It wasn't the Bush precincts.

Of course, had Nader been dropped from the ballot, the whole arrangement would have been corrupted. Who prevented Nader from being removed?

One of the problems with vote-switching of major candidate votes (in addition to the switched-votes counting twice-plus one for Bush and minus one for Kerry) is that they are nearly undetectable.

Your supposition that election-workers were not involved is just that, supposing so. What does the evidence tell you? Someone set this up. Who was it?

[Ms. Toad]

8. I suspect you don't live or vote in Cuyahoga county (or Ohio)

In the place I vote in Ohio, the same 3 precincts have been located together in one polling location for at least 16 years. The polling location even moved from one building to another (and within the same building to a larger room in 2004 to accommodate increased traffic). In those two moves not even as the spatial configuration of the three precincts changed (in all three buildings/rooms I vote in the back right of the room).

I voted in Cuyahoga County for 8 years before that, and the same pattern was true there.

Anytime you have multiple precincts, you should have multiple ballot orders. That's how the ballot rotation is done. Different precinct, different ballot order. Multiple precincts = multiple ballot orders.

As for the locations without multiple ballot orders, I suspect the "Bush" precincts were locations which are prone to be less densely populated (either richer = more land/person or more rural = more land/person) making it more likely that there would be only one precinct (i.e. only one ballot order) assigned per polling location. Since voting is designed to be close to where you live, denser areas tend mean either more polling places located closer together - or grouped precincts in a single polling place. The latter is cheaper, so is what tends to be done.

Precincts are determined by geographical boundaries, which are reviewed and adjusted periodically. I only recall boundary adjustment once during the last 16 years - and certainly not within the period since the previous election. The last record I can find for adjustment of Cuyahoga County precinct boundaries was early 2000. If you live within a particular geographical region you are assigned to that precinct associated with that geographical region. The number of voters within a precinct is the number that lives within the boundaries of that precinct; with respect to the 2004 election that appears to have been set a few years in advance of the election. Similarly, assignment of precincts to a particular polling place on election day and and groupings of precincts generally remain the same from year to year.

Whether it would have taken a genius to arrange or not, to the extent your conclusions require allocating numbers of voters to particular precincts, and numbers of (or specific) precincts to a particular polling place in connection with the 2004 election the reality of how and when these things are determined just doesn't match the assumptions which are necessary to support your conclusions. Data crunching can be extremely helpful, but it needs to be grounded in reality of the circumstances which created the data.

I agree that the skew from a number of factors favored Bush. Voting machines were allocated in a way which resulted in fewer voting machines (and longer lines) in heavily democratic areas. Many decisions which were made by Blackwell in advance of the election tended to disenfranchise poorer (more heavily democratic) voters - and I'll even give you that the vast majority of the voting machine allocation and decision making was calculated to enhance Bush's chances. Folks with less money often find life more difficult, often made that way by people with money. Exercising the right to vote is no exception. As a society we should fix that, and not just the voting disenfranchisement piece of it.

What I disagree with is your conclusion that the vote shifting in Cuyahoga County was by design as part of a scheme to fix the election. Everything I have seen and read indicates it was poor training and/or gross incompetence of pollworkers and poor technology design (which Ohio has been using for years), coupled with systemic accidental or deliberate disenfranchisement of inner city voters (which caused the lines without which the pollworkers wouldn't have been tempted to solve the problem/placate impatient voters by shifting them to an open voting booth).

I am all for addressing problems with elections, which seem to be increasing in number, whatever the source. In this case (assuming Cuyahoga County had not moved to different machines), finding a way to mark the ballots so it can be ensured and verified they are read by a machine which uses the same template that was used to vote the ballot, providing better training for pollworkers so they understand the implications of directing a voter to the wrong booth, and (on the state level) mandating a distribution method for voting machines that ensures that they are evenly distributed by voting population, perhaps taking into account voting turnout history would go a long way toward preventing the kind of screwup that occurred in Cuyahoga County - whether the problem was caused by evil intent, negligence, or just plain incompetence.

I have little patience, however, for the insistence that number crunching proves that all election screw ups are the result of someone or some entity fixing the election. I am interested in the number crunching - it is a field in which I have an advanced degree and I like playing with numbers - so I do look forward to seeing what you have done. But statistics are primarily tools to help manipulate things that are otherwise hard to get a handle on to make the information easier to understand, not solid proof. This is particularly true when the statistics are severed from the reality that gives the raw data meaning. The inflammatory nature of the insistence that statistics prove evildoing - particularly when there are other less inflammatory explanations involving things that can and should be fixed - limits our ability to get those things fixed because the focus shifts to the argument over whether some person/entity did something bad rather that on what can we do to make changes that (even if there is an evildoer) will prevent a repeat of the problems. Anyone/party/entity whose personal reputation is challenged is first going to fiercely defend that reputation, and will be unwilling to acknowledge a problem - let alone be willing to address it - as long as the challenge is couched as an attack on personal/party/corporate integrity. From a practical standpoint, working to fix the acknowledged problems will be a lot easier if we don't insist on placing blame first.

That's a long winded way of saying go ahead and insist it is election fixing if you like. I have addressed some of the basic misunderstandings about how precinct grouping works in Ohio, upon which your conclusions appear to be based, for anyone who is interested. I'll go back to lurking now since I have learned that in this forum it is generally not productive to continue discussions in which I suggest that anything other than evildoing might be the source of the bad election results.

[L. Coyote]

10. I'm reminded of Ronald Reagon's anecdotalism

I'm reminded of Ronald Reagon's anecdotalism with, "In the place I vote in Ohio, the same 3 precincts have been located together in one polling location for at least 16 years.." What of the other 1400+ precincts?

You write, "Anytime you have multiple precincts, you should have multiple ballot orders. That's how the ballot rotation is done."
Not so. It is far more complicated than than you simplification. Multiple ballot orders at a single location is not required to have ballot rotation. Just look at Cuyahoga County; there are locations with multiple precincts and a single ballot order.

You write, "As for the locations without multiple ballot orders, I suspect the "Bush" precincts were locations which are prone to be less densely populated..."

And we all know noone would take advantage of this situation, right?

You write, "What I disagree with is your conclusion that the vote shifting in Cuyahoga County was by design as part of a scheme to fix the election..."

I'm saying is "stinks" and my conclusion is that it should be investigated, not ignored.

You write, "I have little patience, however, for the insistence that number crunching proves that all election screw ups are the result of someone or some entity fixing the election." HELLO. Are you getting this characterization from my few words here? Your argumentation belies an aversion to the facts.

You write, "From a practical standpoint, working to fix the acknowledged problems will be a lot easier if we don't insist on placing blame first." HELLO again. Who is defending whom here. Calling for investigation of vote-switching is placing blame on whom? Who are you defending in advance of blaming someone? Fixing the problem is not made easier with ill-informed denials of the particulars and distraction from the real issue with ad hominem attacks on messengers. That's a transparent, tired old tactic.

[Febble]

12. Oh, well said

I do so agree that if problems are to be solved it is important to figure out what the problem is actually most likely to be.

I shall go back to lurking on the Cuyahoga story....

edited to fix missing word

[Bill Bored]

9. Ms. Toad....

You are obviously a skeptic and that's fine. I don't have time to scrutinize every aspect of Ohio 2004 right now, but the notion of a passive long term subtle predetermined systemic plausibly deniable methodology to favor one candidate over another is very appealing to me. I'd be interested for example in knowing which party it was who pushed for the ballot order rotation amendments back in the 1970s and why, both ostensibly and in reality. It could have been a long term plan to give Republicans a long term advantage. So who pushed it through the legislature 30 years ago? It required a state Constitutional amendment among other laws so someone must have felt pretty strongly abut it. One can only ask why?

Such a plan would also include NOT training poll workers about ballot style encoding, NOT telling them to direct voters to the right precincts, and NOT making sure voters return their punched cards to the right ballot boxes. This would be much better than taking an active role in the heat of Election Day when things, from the point of view of the perpetrators, are more likely to "go wrong."

If I were going to set this up, I'd make sure that as long as nature took its course, votes would on balance be switched in the desired direction, even and ESPECIALLY if the vote switching occurred randomly!

Now my biggest problem with this theory is the way it would become apparent by the excess in third party votes. This limits the number of precincts where it could have occurred. I'd be interested in L. Coyote's estimates of the total amount of switching that could have taken place in the presidential contest in Cuyaghoga. Others have looked at this and not come up with very large numbers, although they were in the thousands of votes and Cuyahoga is of course only one of 88 counties.

Others have come to the conclusion that disenfranchisement or massive deletion of votes in Kerry strongholds could have cost him more votes than switching did, despite the higher efficiency of switching. I hope they see this thread and post here. I just don't have time to do Ohio '04 at the moment.

Does anyone know what ever became of Iceburg? She also had an interest in this and said she had it sussed.

[L. Coyote]

11. Just to really refresh memories

Iceburg was very active with this issue and the original DU work is unraveling what happened in Cuyahoga County:

Chaos in Cuyahoga? 49,000 Votes Disappear into the Ether ...then found?
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x166913

Another person involved was AirAmFan, who wrote, " My theory is that "caterpillar crawl" is associated with long lines and overcrowding, which at least in Cuyahoga were concentrated in poorer areas that favored Kerry. Wherever the polls are overcrowded, the candidate most people favor is likely to lose votes to "caterpillar crawl". From what I've heard, places that favored Bush generally were not overcrowded and often had precinct staffers who accompanied each voter to a booth to make sure it was the right one and that the punchcard was inserted correctly."

jmknapp was also very productive. His maps were excellent.

Here is another thread to refresh memories of who was active:
"Uncounted Votes in Cuyahoga County" by R. H. Phillips. How many??
http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=196518

That thread links to a lot of the previous discussions and to Joe Knapp's maps on Daily Kos.