You would think that just one of the “researchers” could have researched their topic a bit further. There is plenty of available literature PROVING the irregularities that occurred with the punch cards in the counties of interest in 1996 and 2000. If one studies the spoiled ballots numbers in those years that could have been caused by democratic voter fraud by putting a single additional punch on otherwise valid Bush/Dole votes, while similar mispunches did not occur on anywhere near the same number of Gore/Clinton votes, nor did they occur in any similar numbers for other offices - amazing.

I guess our Berkeley researchers simply could not read existing books and articles showing definitive democratic vote fraud in 96 and 2000, nor could they even consider that the skew could have been in the election 2000 numbers. Therefore when a much better system was put in place - it worked to help eliminate the past voter fraud in those counties!!!

Even a few lone voices on the DU site pointed out some of these ideas/published data, but they were shouted down by our rabid tinfoil researchers!

Lonewacko - you seem determined to prove that the voting system in the US is getting worse, when in fact it is getting much better. People who create doubt where there is hope are the ones who disenfrachise voters. The facts simple are that since 1960 the voting system has continued to improve and will continue to improve, but sore losers are convincing more people that the system is not fair. Positive changes are good, they happen through involvement locally, not behind a keyboard doing bad science/bad ststistics and passing it through the blow-holes of the liberal media establishment.

Richard Nixon likely won the 1960 election, but knew that exposing the horrible underbelly of vote fraud would do more damage to the US than it could possibly do good. Improvements have been tremendous and more needs to be done, but not by destroying the confidence in the system. There are bad calls in every sporting event, mismarked prices in every store - Americans are not perfect, but America is the BEST. Stand by the country and respect the leaders.

Posted by: I collect political items at November 18, 2004 10:37 PM

The three counties where the voting anomalies were most prevalent were also the most heavily Democratic: Broward, Palm Beach and Miami-Dade, respectively

Gee I guess the rest of FL turned out by the 100 thousands to reelect President Bush to be sure the snowbirds that are registered in FL and NY didn’t get a chance to decide for FL.

Posted by: RL at November 18, 2004 11:56 PM

The headline for this piece could just as easily be “Bush Cheated out of Hundreds of Thousands of Florida Votes in 2000.”

This study is esentially comparing statistically 2004 to 2000 and extrapolating votes based on the “anomaly.” But the headline above reflects the presumption that the 2000 counts are accurate, in spite of those votes being cast using known inferior paper technology, versus the known superior accuracy of electronic voting. If we presume that the 2004 vote was the accurate vote, then we get the headline I suggest above.

This study is just an attempted smear of electronic voting. The leftists want paper receipts so when they lose close elections their lawyers have a way to hold the election over and over until they “win.”

:jackson

Posted by: jackson zed at November 19, 2004 03:35 PM

While I am in complete agreement that this study is badly flawed, and that Bush is the winner, the claim that only “leftists” want paper receipts is flawed.
Name me as one conservative who wants any electronic voting system to be required to produce a tangible record of each vote cast, not merely an aggregate electronic total.
Something along the lines of a cash register roll receipt holding an appreviated record of each vote cast. These rolls should ideally be transported and stored separately from the electronic records, and be made available for both spot-checking and, if need be, full forensic recovery of votes cast.

Posted by: j.pickens at November 19, 2004 10:42 PM

appreviated = abbreviated sorry…

Posted by: j.pickens at November 19, 2004 10:43 PM

jackson….I don’t get you at all. No one needs to “smear” electronic voting; it’s record is clearly lacking. There hasn’t been a single pristine electronic vote. Now, while I accept that one day electronic voting might be a valid option, by no one’s imagination is that day at hand. Why you would object to something as simple as the equivilent of an atm receipt I don’t grasp; what is your point?

Posted by: Jatsby at November 19, 2004 10:56 PM

j.pickens!

Name me as one conservative who wants any electronic voting system to be required to produce a tangible record of each vote cast, not merely an aggregate electronic total.

From what I understand, most of the electronic systems in use (including the Diebold systems) do exactly this; they don’t just record vote totals, but store in several redundant memories each ballot “event,” which in turn can be printed out.

Where I live, we’ve been using electronic voting for years, as have other parts of the country. Nobody ever complained about it until after the 2000 election and voting reform starting being discussed on a national level. I don’t believe that it’s a coincidence that
the people who were bitching about “paper trails” before the election are the same folks running around talking about how Kerry “really” won the “stolen” election of 2004. And although no one has any way of really knowing, I can’t shake the feeling that if Kerry had won, we wouldn’t be hearing a peep out them now, nor would we be having this conversation.

On the topic of paper trails, I’m personally agnostic; what I don’t want is a way to “do-over” elections when one side or the other doesn’t like the result. The left, demonstrating their customary myopia (see: special proscecutor law) doesn’t seem to care about this at all. To them, a system that doesn’t produce the outcome they desire is broken by definition.

The Founders of this country doled out the responsibility for determining the “time, manner, and place” of elections to the state legislatures in order to keep the process as close to the people as possible and also because they knew that X number of decentralized systems would be far more difficult to co-opt than a single, monolithic system. I remain convinced that the best system will be had by maintaining the Founders’ vision of separate, dispersed, locally accountable systems, all coming up with competing ideas for the best way to conduct their balloting.

Jatsby!

jackson….I don’t get you at all. No one needs to “smear” electronic voting; it’s record is clearly lacking. There hasn’t been a single pristine electronic vote.

Nor will there ever be. Elections will always be constrained by nature, and thus will always incur a certain amount of variation. Perfection simply isn’t one of our options, so to argue against electronic voting because it isn’t perfect isn’t really an argument at all.

Although these electronic systems are just now being put into widespread use, I’ve yet to see any serious—- much less convincing—- argument that these systems aren’t in general more accurate than the paper-based analog systems that they’re replacing.

Errors produced by analog systems are most likely to fail on individual ballots, making the error they produce more or less undetectable, whereas electronic voting machines are most likely to fail at the machine level, making the errors produced far larger and due to their nature, more likely to be caught.

:jackson

Posted by: jackson zed at November 20, 2004 03:56 PM

If you look at Figure 1 in the paper, you’ll notice that the actuals in the upper end of the scale for % Democrat Support Election 2000 are all below the estimates for e-voting and all above the estimates for non e-voting. In addition, the actuals at the lower end are all below the estimate for non e-voting. These observations indicate that there is a serious model specification problem for the e-voting counties and a biased parameter estimate for the non e-voting counties.

This paper probably would not pass muster in the peer review process.

T. F. McDonough, Ph.D.

Posted by: DrMack at November 23, 2004 02:44 PM

The Berkely study on the presidential vote patterns has a couple of problems. They start with the assumption that there is a problem with the electronic voting.

1. The theory you start with and the methods you utilize can determine what you find.

2. In regression and related regression models, what you leave out is sometimes more important than what you put in. I do not recall seeing any variables addressing orthodox jewish voters (more likely to vote for Bush in 2004), who are an important block in Broward, Palm Beach and Dade counties. Also, Lieberman was on the ticket in 2000. Futher, it may be that Bush did better with older voters than in 2000, not to mention pro military anti terror Democrats.

Anyway they have left out plausible explanations for the increased Republican vote in those counties and only examined variables that may support thier original theory of vote fraud.

Further, regression and regression related approaches are problematic when you use a lot of correlated variables. The betas tend to be less than stable and the signs may even reverse. In other words do not bet your life on the results.

Posted by: Bugoff at November 24, 2004 03:58 AM

Quoth DrMack:

This paper probably would not pass muster in the peer review process.

But remember, this is Berkeley. Political agenda trumps academic integrity. And not just at the school.

Posted by: gus3 at November 27, 2004 01:33 PM

And we all know Berkeley is a bastion of non-partisanship.

Posted by: Oyster at November 28, 2004 09:01 PM