« [fun with] The Biography of George W. Bush | Main | Bev Harris Finds Official Vote Records in Trash »

November 16, 2004



You're taking the average of the last four elections for each county, and calling it a trend. It's not a trend.

Two of those elections had Clinton winning Ohio. That's what pulls the average up.

All you've shown is that some voters voted for Clinton that didn't vote for Kerry. That's not irrefutable evidence that the ohio election was rigged.

ledge of liberty


thank you for your comment, but you're incorrect [that sounds rude, but it's not meant to be]. this is based on an average of the past 4 elections. That includes two wins for the republican party and two wins for the democrat party. the best way to ascertain the likely voter is to use data proving who they voted for in previous elections... and that is what you have in this report. for instance, several of my projections for certain counties are dead on, within only 20-30 votes off from the actual 2004 results. this should be the norm, but then you have a county like mercer that would have had to have taken almost 30% of the democrat votes in the official tallies... it just doeesn't add up.

You have to take a closer look at the data presented -- not just the bottom line.


ledge of liberty

due to some controversy over at the daily kos regarding the title of this story, i've changed it. typically this wouldn't bother me and i wouldn't make changes, but i want people to download the data i've put together and i don't want a silly thing like a title to stop them. so...r.i.p. "irrefutable evidence" and hello "statistical evidence". [though i still think it's irrefutable, so there :P]



try comparing '88 to '92,

'88 + '92 to '96,
'88 + '92 + '96 to '00.

You'll see variance in every election. It's what happens when a party does better than he did over an earlier period of time.

If a county has bigger variance than another county, it can just be explained by higher turnout from a demographic that favored one candidate.

ledge of liberty

yeah, i totally agree there'll be variance, but they'll be small campared to the 2004 election. 1%, 2%, 3%, 4%... okay, sure... but 27% of the vote -- no way is that legit.


You also leave out the Perot factor in 1992. Probably the best and only comparison for a two-man race would be 1988, and the numbers don't look so startling. Basically, all this proves is that there is up to 20% of the population that can be called "swing voters". Quel choque!

ledge of liberty

hahaha... i was waiting for someone to bring up the swing voter aspect. okay, first off, 20% of the population are not swing voters. the swing vote represents a very small amount of the vote... just enough to swing an election -- hence the name. touche!


Where did you study statistics? I don't have time to detail all of the falacies in your methodology. Suffice it to say very simply, your own data says that 51.8% of Ohio voters voted for Bush in 2000. 51.3% voted for Bush in 2004. Sounds fairly consistant to me. Figures lie and liars figure! Find someting constructive to do.


I don't disagree with your methodology, but even under your "Projected Votes" column which i suppose is what the vote would be if there were no "tampering," Bush still wins by 141,000 votes. So how does this help?


Dude, are you still defending this study? Go do the alternate studies I suggested. The fact is that it's completely possible for there to be huge variances.

Look, say that A beats B by 15%. Then A beats B by 5%. Then B beats A by 5%.

For the fourth election, it wouldn't be weird to consider B beating A by 15%. It would just be part of a trend. But that would be a variance of what, 20%? If you compared it to the previous three?

ledge of liberty

Wow... are you people being annoying or what?

Dave wants to solely compare the election to 2000 and that's it -- where did you study statistics? kindergarten?

Dan [who's not annoying], you have to take the average percent change and apply it to the total # of votes.

Tunesmith, you don't seem to be able to wrap your head around the math, and I don't think you will. What you are proposing isn't how I came to my conclusions. I suggest, if you really want to understand this, you sit down and read the article and actually attempt to understand the math behind it.

This is highschool math people -- nothing too complicated or abstract.


Peter, you're killing me.

Look. I'm trying to explain to the difference between an average and a trend. If you are better at statistics than me, you should know the difference, right?

Plot these four points at regular intervals:

-50, -10, 30

That's an exact gain of forty points each time, right? You'd expect the next number to be 70.

Now, look at a second case.

-7, -3, 1

That's an exact gain of four points each time. You'd expect the next number to be 5.


Those are trends. If the first case really did end up with 70, and the second case really did end up with 5, it would be exactly in line with expectations.

Now, let's do what you did. Let's average the numbers. In the first case, the average is -10. In the second, the average is -3.

In the first case, if you got 70 points instead of -10, you'd be off by a whole 80 points. In the second case, you'd only be off by 8 points. The first case has TEN TIMES THE VARIANCE as the second case.

It proves nothing.

Let's look at one of your actual examples. One of your counties. Hell, let's look at Mercer, the one that you're most alarmed about.

First, we'll compare '92 to '88. Taking an average of one value is by definition that value. It's kind of silly, but not any sillier than what you'r'e doing. In '92, you'd expect the votes to be the same as '88.

Dem: variance of -1.9%
Rep: variance of -22.2%

Wow, Republicans down by 22% ! Fraud?! Well, no. Just a lot of Perot voters.

Now we'll compare '96 to the average of '88 and '92.

Dem: variance of 27.8% !!
Rep: variance of -11%

Wow, check out that variance. Did Clinton commit Fraud to beat Dole? Well, no. It was just a lot of people voting to re-elect Clinton who didn't vote for him before. Plus, Perot was still out there drawing votes from the Republicans.

Now we'll compare '00 to the average of '88, '92, and '96

Dem: variance of -3.2%
Rep: variance of 30.6% !!!!

Wow, 30%. Does that mean Bush committed Fraud to beat Gore in Ohio? Well, no. There were just a lot less third party votes taking votes from the Republican, and Gore was less attracted to conservatives than Clinton was.

Look, all three years had variances of the same (or even greater) than the greatest variance in your spreadsheet this year. It proves nothing! Everyone knows Bush was successful in turning out a lot of conservatives who hadn't voted in previous elections.

Peter, this is a horrible study and you're embarrassing yourself when you tell the rest of us we don't know what you're talking about. I mean, go look at cell AC92. You're not even taking a weighted average, Peter. If I'm in a three-person county with my parents and they vote GOP and I vote Dem, the average for that county is 33% Dem. If the next county over has 25,000 people and goes 50/50, you can't average the percentages together and say the average Dem/Rep breakdown between the two was 58% GOP and 42% Democrat.

Peter, this is embarrassing. If you care about your credibility, you'd withdraw this study ASAP.

ledge of liberty

First off, about "Peter, this is embarrassing. If you care about your credibility, you'd withdraw this study ASAP."

This is a BLOG on the INTERNET... get a grip.

Secondly, if you can step out of your remedial math bubble for a moment, look at my projected votes, you'd see that about 50% of my projections are dead on. Some of my projections come within 20 votes of the actual results. A 27% change is evidence of fraud.

I can't hold your hand on this any longer, if you haven't gotten it by now -- you never will. Therefore, I'm done with you...

Thank you, drive through. Next!



Tunesmith is completely dead wrong. I looked at the study and found it to be well researched, formulated, and insightful. The numbers make perfect sense if you understand the math behind it.

And after looking at Tunesmith's website (every other post is how there was no election fraud or is a debunk of some exit poll study) it doesn't surprise me that he/she refuses to acknowledge your study as revelvant. just ignore...

Good work Peter! Thank you and keep it up!


I'm no statistician, but it just seems odd to me, from a purely common-sense point of view, that almost all the "glitches" and "discrepancies" that have come to light since 11/2 have fallen in Bush's favor. If these were truly random errors, it seems that Kerry would have benefited at least occasionally. That doesn't seem to be the case in this election, which leads me to conclude that, more likely than not, this election was stolen.


I'm really trying to see your point Tunesmith, but it's just not there. This study shows that up to 27% of the Democrat vote is just plain missing from certain counties, while other Counties have remained completely unaffected. You're sighting mathmatical examples to prove your point, but it doesn't match up with how the math for the study was actually done... and I'm no mathmatician, but it all seems quite logical to me. Thanks Peter!


I'm too dumb to figure out the math, but I did think of a couple of things, and I was wondering what you thought about it:

1. Did you check to see what the population increase of eligible voters was for the past 4 years, and (if possible) what their political leanings were?

2. You seem to be assuming that a person votes for a political party much like they belong to a blood clan, i.e. if you're registered as Democrat, you vote Democrat, & vice-versa. Could it be possible that Bush's assault of propaganda and fear-mongering actually worked? I mean, it was false advertising to be sure, and I suppose you could say that in a way, that's "stealing", but don't you think that in the face of the media assault, a large number of democrats changed their mind and votede Bush?

ledge of liberty

hi anonymouse -- to answer your questions...

1. these numbers were created by applying a percentage to the actual 2004 votes -- i didn't adjust for population. there is a chance this could affect a percentage point, maybe two -- so if anyone would like to adjust my numbers for new and likely voters -- please go right ahead. i'll post the results here. but, as far as i can find, voters party affiliations are undeclared in ohio.

2. the first sentence of this story is -- "To believe Bush won Ohio, you would have to believe that perfectly good Democrats turned Republican." if one doesn't agree with that, they might as well stop there. personally [and obviously], i don't think that 10-30% of democrats in certain counties decided to vote bush this year. i think that's nonsense -- it doesn't reflect the polls, the exit polls, the zogby projections, the polarization, the history, or anything else.


my friends; in my opinion I do honestly believe that Bush stole the elections. For days afterwards I pored over pre-election polls, exit polls, election results, etc.
I do believe that there was serious, massive voter fraud in Ohio and Florida, in particular.
HOWEVER, I have also not seen any HARD EVIDENCE.
Unless you have programmers from Dieboldt come forward and admit to doing it, and they turn out to be credible to most of the American pubilc, it's a FUCKING MOOT POINT!! Let it go!
Concern yourselfves more with ending the war in Iraq; something positive and tangible that will help peoples' lives.
There's no real hard proof of voter fraud; I am sorry....


Some will argue that we all should get over it and move on. That is not the strategy of Bev Harris (www.blackboxvoting.org) who is in Florida collecting hard evidence that polling tape records are being destroyed and forged polling tape records are being substituted. Recounts in New Hampshire, Ohio, and Florida may well show hard evidence of election fraud if the FBI, GAO, and the DOJ's Voting Rights section do not manage to run out the clock while evidence is destroyed and forgeries are substituted. Let's demand fair, nonpartisan recounts and join Common Cause in petitioning that Kenneth Blackwell recuse himself from participating in any recount because he has shown himself to be blatantly partisan (http://www.commoncause.org/DemandPureOhioRecount).


In addition to Peter's analysis, I compared the 2000 election to averages from the elections going back to 1988, and I compared the 2000 election and the 2004 election.

I found that (1) the 2000 election departed from the three previous elections MORE than the 2004 election departed from the four previous elections, and (2) the 2004 election departed from the 2000 election MUCH, MUCH LESS than the 2004 election departed from the four previous elections.

My analysis suggests that it was the 2000 election that broke from the "trend", and that the 2004 election is largely consistent with the 2000.

I see no reason whatsoever for treating the 2000 comparison of Dem. and Rep. percentages as less valid than other years.

Further, we do have a reason for treating the 1992 comparison of Dem. and Rep. percentages as less valid than other years. Why? We are just comparing Dem. and Rep. votes to one another and ignoring third parties. In 1992, Perot was a very significant third party candidate who decreased the number of Rep. votes relative to the number of Dem. votes. Perot got 21% of the vote in ohio in 1992.

Then Clinton ran as an incumbant in 1996. In 2000, Dems did not have the benefit of a right of center third party or incumbancy. Hence, it was the 2000 election that broke from the "trend."

I want to pinpoint any and all places where fraud seems most likely. No offense to Peter, but this effort is not helped by faulty analysis.


Tunesmith has done what many have done and that is debunk the "efforts"of the whining sour grapes Dems to find ANYTHING they can to avoid another humiliating loss and/or psychpotherapy. When Tune completely unravels the analysis, which was quite simple to do, Peter washes his hands of the infidel....LOL. Nice try Peter but you can lie with stats ANYTIME to suit your needs, it was a class in college. Moveon.org with life my friend. We do have some wonderful parting gifts for you though.


What most self-proclaimed debunkers overlook is that they are shooting themselves in the foot. Whatever shall Neo "Conservatives” do when they need to claim they were cheated at the polls? They will do what their opponents are, but if they are successful at “debunking” facts now, that will work against them later.

Another thing, over time red becomes blue, blue becomes red. Age? Income? Who knows why... but once the middle class is completely removed, there is nothing to protect the relatively small "upper class" because by then they will have made enemies of the lower class. Lower and middle class Neo "Conservatives" are truly cutting their noses to spite their faces.

Red or blue will lose a defining distinction and it should as things stand. The two major parties have been invaded by Neo "Conservatives" which accounts for mass confusion and breakdown within both parties. Democrats and Republicans alike have been wondering what is going on with their lawmakers and representatives. Watch as they expose themselves though the how and why of what they support.

* People can only be pushed so far.

* Election analysis is the evidential beginning of a revolutionary movement.

* Lawmakers contrive not to be victims of their own crimes, but voting irregularities are backfiring without respect to initiators. Criminal charges have been filed against election officials, with more on the way.

Ignorant debunkers of voting irregularities should pick a side with the future in mind. Voting irregularities have been notable since the 1930s and become common place these last several decades. A very real majority of people feel breathtaking real time squeezes, and without legitimate voting we are left without a republic.

To others who have performed analysis for Ohio and other states, please forward those and precinct data files (if available) to election.statistics@gmail.com for viewing or further analysis.

Thank you Peter, for the significant time and effort you chose to act with through your analysis. Thank you also for putting it "out there" for review and comparison. -- CS Colvin


Posted by: straw_citizen:

"Ignorant debunkers of voting irregularities should pick a side with the future in mind. Voting irregularities have been notable since the 1930s and become common place these last several decades. A very real majority of people feel breathtaking real time squeezes, and without legitimate voting we are left without a republic."

Yes and I am sure that is why the left is going through the whining and all out effort to delegitimize, at all cost, the Bush win in November. If Kerry won by the same margin the story is over and done with and you all know it. GET OVER IT and join the Americans who support the Pres, almost 61 million of them. Oh and Merry Christmas!!


Somehow an evil man became president by scaring and tricking half the people in this country...the figures don't really matter, but I agree there was fraud involved in the election.

The comments to this entry are closed.