Friedman: Democracy in America Has Officially Become a Privatized Circus

[AtLiberty]

Democracy in America Has Officially Become a Privatized Circus

Brad Friedman
Huffington Post
11/26/05

Pollster Mark Blumenthal Joins the Crowd of Folks Who Simply Don’t Get It or Just Don’t Care

Pollster Mark Blumenthal follows up his recent post on the matter of the extraordinarily questionable results of Ohio’s November 8th, ‘05 Elections. All four of the Election Reform initiatives on the ballot that day mysteriously and spectacularly failed to win approval in stark opposition to the pre-election polling by the historically accurate Columbus Dispatch which had predicted most of them would win by large margins.

That they lost by large margins instead, has been the subject of some controversy. We’ve blogged about it (here and here) and responded (here) to Blumenthal’s analysis of tweaks in the Dispatch polls methodology that he feels “might” have skewed their findings this time around. Despite the major changes in the “methodology” used to gather the actual Election Results in Ohio this year (44 of Ohio’s 88 counties installed all-new Touch-Screen Electronic Voting Machines for the first time ever) Blumenthal’s presumption is that it must have been the tweaks in the Dispatch poll’s methodology rather than any problem with the Election Results as gathered by these all-new, untested, unaccountable, untransparent and demonstrably prone-to-failure Electronic Voting Systems...

http://www.huffingtonpost.com/brad-friedman/democracy-in-america-has-_b_11253.html

[Hand]

1. Ha! Thought for a second it was THOMAS Friedman...

Fat chance.

[Doctor_J]

2. Anyone who believes we're still living in a democracy

probably also believes that Limpballs telles the truth and that George Bush is a man og God. Time to turn off Limpballs and get some real information piped in.

[Kolesar]

3. Fraud caused a double-digit % vote swing--Bullshit--eom

[Bill Bored]

4. Well we've heard here on DU that these initiatives were just too complex

for voters to comprehend from reading the ballot, or that even though they agreed with them, they just didn't trust their government to actually implement them (I'm open to suggestions as to who else might be able to do so), or that the polls were conducted too far in advance of the election to mean anything. Something like that anyway.

So if anyone wants to try to debunk any of this stuff, which all sounds perfectly reasonable to me, please go ahead.

We should try to get to the bottom of this as the discrepancies were major ones this time and the election results for one of the issues on the ballot matched the polls quite closely. That's weird.

It might also be helpful if some of the pollster gurus would take the time and trouble to learn a little bit more about how e-voting works. But I guess everyone has their respective areas of expertise. Far be it from me to suggest a more eclectic approach to this stuff.

After all, it's just control of the world's most powerful democracy and sole superpower at stake here. Nothing major like how the soap should smell and who's going to buy it.

So just keep on polling guys and gals! Maybe in a few decades or so, there will be enough statistically significant discrepancies between the election results and the polls for you to think twice about what's been going on. We can all afford to wait that long, can't we?

[Febble]

5. Well, I was pretty openminded

still am, but I do wish that instead of lambasting those who try to look sensibly at the evidence, people would, well, look at the evidence.

We know polls suck (or I do).

We know the election system sucks (especially in Ohio).

And it looks as though, once again, at least one of them was wrong in 2005.

MysteryPollster has identified some things that might have gone wrong with the polls. Others have pointed out that the glitch coincides with the extension of DRE voting.

So what we should be doing is constructing a decent hypothesis to try and distinguish between these two. And the obvious one is: if the discrepancy was due to DREs, the discrepancy should be higher in counties with DREs. No?

OK, Philip Klinkner's done it, here:

http://polysigh.blogspot.com/2005/11/fraud-in-ohio-doubtful.html

He did a multiple regression and found that voting technology was not significantly associated with the magnitude of the discrepancy between poll and count. So it looks as though either the fraud was pretty uniform (which rather undermines the popular hypothesis), or Ms. Toad, TheBorealAvenger, MysteryPollster and others may be on the right lines.

I'd like to know what Friedman thinks of Klinkner's analysis.

[Stevepol]

6. Febble, may I ask what accounts for your interest in American exit polling

I know you are a British or Scottish legislator and yet you seem to spend an inordinate amount of time on DU debunking anybody in the US who would suggest that a vote such as this one in OH, where the discrepancy was around 40% between the pre-election poll (which as one of the articles you refer to indicates has always been reliable in the past) and the alleged electronic tabulation result might be fraudulent.

Things must be dragging in the UK I take it.

Where's your buddy OTOH? He usually is hyper-alert to debunk as well whenever you post.

Cheers, me lassie.

[Febble]

7. I'm interested

in the conduct of American democracy because it matters to the rest of the world, sadly.

I was interested in the 2004 exit poll story because I suspected that the exit polls may have been right, and that the count was corrupted. I had desperately hoped that Kerry would win. I regard it as a tragedy for more than just Americans that he was not inaugurated president.

I still think that there is massive evidence that the 2004 election was corrupt.

But I have not, for some time, believed that the exit polls support the case that there was massive electronic vote-switching. Indeed, I think that the evidence suggests that there was not.

And I therefore think that arguments based on the exit polls detract from the validity of the case.

As far as the recent Ohio referendum was concerned, I was open minded. But what I hoped someone would do would be to do a multiple regression to see whether the poll-count discrepancy was indeed greater in DRE counties. Klinkner has done this and found it was not.

I'm trained to look at evidence objectively. I try to do so. Klinkner's analysis is part of the evidence, as is Mark Blumenthal's.

[Stevepol]

9. Maybe you could enlighten me as to what made you change your mind.

You say the reason you are interested in this exit polling stuff is that you "suspected that the exit polls may have been right, and that the count was corrupted."

Then a paragraph later, you say that you "have not, for some time, believed that the exit polls support the case that there was massive electronic vote-switching. Indeed, I think that the evidence suggests that there was not."

What made you decide that your suspicions were unjustified? The work of the Election Archive people, Steve Freeman et al.? The number crunching of TIA?

I might ask as well, what level of discrepancy would be required before you would think there might be something to your suspicion that "the exit polls may have been right"? From your other posts it has seemed to me you don't believe in exit polls period. Am I wrong?

[Febble]

12. Well, clearly

the work of the people you cite argue the other way, so no, not those people.

And I want to make it clear that I do not argue that there was no fraud - there are plenty of ways of stealing an election that won't show up in the exit polls anyway, including voter suppression, but also some forms of vote tampering.

But I do think the evidence now weighs fairly heavily against, specifically, large-scale vote switching (i.e. vote switching on a scale to account for the exit poll discrepancy - even a relatively small proportion of it). I laid out this view at some length here

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=398267&mesg_id=398267

so I won't repeat it all again.

But in answer to the questions in your last paragraph: it's not the level of a discrepancy that would convince me, but the extent to which it can be "accounted for" (statistically), or not, by convincing factors. For example, had the discrepancy in the Ohio referendum been statistically greater in counties with DRRs I wouldhave found that fairly persuasive. It was the first thing I thought should be investigated. As Bill Bored says, the null finding from Klinkner's analysis does not rule out fraud, but it infirms one specific (and popular) hypothesis.

I don't believe that exit polls are a very good way to audit an election, although I agree they seem to be all you've got right now. I certainly think that the 2004 exit polls were not designed to do that job, and therefore do it particularly badly. A better designed poll, designed specifically as a check on the election, might work. Ironically Edison-Mitofsky recently conducted such a poll in Azerbaijan.

But I do think it is important for people to be aware that the "Margin of Error" computed for any poll is the Margin of Error assuming random sampling. No poll can actually assume that, as non-sampling error is a major problem for any survey. Unfortunately you can't calculate it. However, you can examine the methodology fairly carefully, and some polls are simply better than others. The BYU exit poll in Utah was an example of an extremely well designed exit poll. It also got neared to the count in Utah than did the E-M poll.

[Febble]

8. BTW

I'm certainly not a legislator. I'm a research psychologist, and therefore trained in statistical hypothesis testing.

I tested a few hypotheses here:

http://uscountvotes.org/ucvAnalysis/OH/FranklinCountyReport_v2.pdf

They don't all come out the wrong way ;)

[Bill Bored]

10. Unfortunately, this doesn't prove much.

In the case of Ohio, there are so many ways to steal votes, it doesn't particularly impress me that Diebold and other DREs don't stick out like sore thumbs compared to other voting methods.

That said, at least they don't have ballot order rotation for referendums the way they do for candidate races, so that's one thing they can't screw around with.

But we still have the old "central tabulator" stuff, and the VVPATs that come out blank because someone put the paper in the printers backwards in Lucas County, and Republicans in charge of the electoral process and THEIR agenda winning at the polls.

I guess I just find the reasons for "normal" people voting No to be a bit vague and unconvincing. Sort of like voting for the Shrub.

I suppose there will be other opportunities for reform besides these constitutional amendments and frankly, I'm not sure why they had to go that particular route in the first place.

And I still think that if one doesn't understand a ballot question, they would be more likely to abstain than to vote No. But then again, I'm not a poli-sci guy.

Speaking of which, I wonder what a poli-sci guy would consider to be bona fide evidence of election fraud anyway, assuming it's not some job security sort of thing to NEVER consider ANYTHING! After all, if all our elections were rigged, would there still be a need for Poli Sci? Or polls? Or any of that stuff? It would just be window dressing at that point, wouldn't it? Just askin'.

[Febble]

11. Good point

but it takes us further to know that if there was fraud, it wasn't markedly machine-related.

As for what would convince a political scientist about fraud: have you read Mebane on Ohio 2004 or Florida 2000?

[Bill Bored]

14. As I understand Mebane's Ohio stuff

he was looking for a lack of correlation between some 2004 races and between 2004 and 2002 or 2000 races at the precinct level, right? And instead he found they were correlated.

So if they were not correlated, that would convince him there was fraud???

Why couldn't one just argue that people simply changed their minds since the last election, or that they wanted to split their vote between the parties for the President and the Congress or the Governor a la, "Republicans for Kerry", etc?

Personally, I think the fact that they WERE correlated is more of a red flag, given Bush's disgraceful performance and also the reported increases in Democratic registration in Ohio.

I think if we try hard enough, we can come up with an alternate explanation for just about anything, can't we? And the poli-sci folks for better or worse, seem to be masters at this.

I'm not saying they're necessarily wrong, or evil, or anything like that, but since there is no basis for confidence in our election system, they should not be dismissing those such as Brad, who question results, as just "tin foil hat" wearers.

And I'm also suggesting that the poli sci folks do have some investment in the status quo. What would they have been doing in say, the former Soviet Union? If they were allowed to exist at all, I'd think they would be primarily engaged in propaganda. If our elections really ARE rigged, and these folks keep saying everything is fine, how far away are we from that totalitarian model then?

Again, I'm just askin'. Some of best friends took Poli Sci!

What else ya got?

[Febble]

15. Well I was thinking specifically

of his paper on machine allocation in Franklin County, which was a much more detailed and sophisticated analysis than my own:

http://macht.arts.cornell.edu/wrm1/franklin2.pdf

Conclusion
The allocation of voting machines in Franklin County was clearly biased against voters in precincts with high proportions of African Americans when measured using the standard of the November, 2004, electorate. In precincts with high proportions of African American voters there were 13.6 percent more active voters per voting machine than in precincts having low proportions of African American voters. While shortages of voting machines caused long delays in voting throughout the county, the allocation of voting machines among the county's precincts affected different voters differently. The most severe effects in terms of reduced voter turnout were incident on voters in precincts that had high proportions of African Americans. The most conservative estimate—based on the reported size of the active electorate in November—is that typically the shortages of machines reduced voter turnout by slightly more than four percent in precincts in which high proportions of the voters were African American, while shortages in precincts where very few voters were African American reduced voter turnout by slightly less than 1.5 percent.

If the allocation of voting machines is compared to information about the size of the active electorate that was available to Franklin County election ofcials at the end of April, 2004, then the allocation of machines is not biased against voters who were active at that time in precincts having high proportions of African Americans. But if election ofcials did use that information too make their allocation plans, then they made plans that involved using a total number of machines that was nearly 45 percent too small. Even using the April measure of the size of the active electorate, 5,023 working voting machines were needed, not 2,800 machines as data supplied by
the county indicate were actually deployed on election day. And sticking with plans possibly made using the information from April meant that the ofcials ignored information during thee summer and fall that showed that the November electorate would be substantially larger. Between April and November, voter registration in the county increased by 15 percent. If nothing else, the surge of new registrants should have been a clear indicator that plans made based on the April information would prove woefully insufficient.

Also reported on TomPaine:

http://www.tompaine.com/articles/20050721/timing_and_turnout_in_ohio.php

And I don't know whether Mebane has dismissed anyone as a tinfoil hat wearer, but he did write a paper called "The Wrong Man is President" - (re Gore/Bush) which hardly suggests an investment in the status quo.

[Bill Bored]

17. OK, but he's not the one maligning Brad on Polisigh (Klinkner)

Has anyone actually checked to see if there were even that many machines in existence to be used on election day? It can't be "fraud" if there weren't any machines to be had, can it?

But he does seem to be intimating that at least there was negligence, so I guess his head is above the sand.

[OnTheOtherHand]

18. Mebane on Ohio 2004 (on fraud, not machine allocation)

It would be strange to expect no correlation between 2004 and previous results, even in the presence of substantial fraud. And anyway, it was immediately obvious on Election Night that the results were correlated -- it isn't as if Cuyahoga County went Republican. Mebane was looking for evidence of some other pattern beyond that -- for instance, that a bunch of touch-screen counties were way out of line, or that results in 15% of the precincts seemed to be out of line with the others as if Bush had been given a 10% boost in those precincts alone. (Nothing in statistical analysis is biased toward making abnormal things look common; statistical analysis confirmed that the "Jews for Buchanan" vote in Florida 2000 was bizarre.)

The easiest way to evade Mebane-style analysis would be to control the entire vote count. If, say, Blackwell had the ability to redo the entire shebang from his desk, then he could just sit back and snicker while people chased red herrings like the thousands of votes in that Gahanna precinct. My general point here is that fraud is easier to conceal if it is very widely distributed, rather than concentrated in a particular kind of place.

I want to reiterate my first point. You wrote, "Personally, I think the fact that they WERE correlated is more of a red flag, given Bush's disgraceful performance and also the reported increases in Democratic registration in Ohio." But correlation doesn't mean that Bush does about the same in 2004 as he does in 2000; it means (since we are talking about positive correlation) that the places where he did best in 2004 tend to be places where he also did best in 2000. If Kerry did no better in (traditionally) Democratic precincts than in Republican precincts, then we would all be screaming.

Just as an aside: There were political scientists in the former Soviet Union, so we don't have to speculate about that as a counterfactual. They came in about the same proportion of gutless and courageous as any other profession, as far as I can tell. But as a matter of historical record, many younger Soviet political scientists supported democratic reform when that was far from being a safe career move. As a young fry just out of college, I got to work with some folks who fielded a public opinion poll that criticized Soviet foreign policy. In retrospect or in the U.S. perspective it seems pretty tame, but at the time it was a calculated risk.

I have seen two moderately detailed inquiries into evidence about the 2005 Ohio issue results, but I don't know which "poli-sci folks" have been "dismissing those such as Brad, who question results, as just 'tin foil hat' wearers" or "saying everything is fine." Anaxarchos has appropriately cited Klinkner as an expert on how ballot spoilage has disenfranchised black voters -- not exactly a Panglossian establishment hack. The poor guy can hardly help it if he was curious to see whether there really was strong evidence of fraud in Ohio 2005, or that his particular test came back negative. He could have chosen not to report the result, but that would have been a hack choice. I see no reason to pick a fight with him over this.

See ya....

[OnTheOtherHand]

19. oh, OK, I was reading the wrong Klinkner post

Klinkner's first post was essentially informative. His second one arguably offered practical advice. And then he got snarky.

Imagine that. Snarky comments on election fraud. ;)

_And_ two links to transcripts of Bob Fitrakis interviewing Lyndon LaRouche. Ouch.

Brad Friedman should pick his enemies more carefully, IMHO, but then, a lot of us should. Anyway, sorry for my confusion about the context.

[onthebench]

13. Which VVPAT was used in OH?

There were none certified at the time?

Since the Do Not call was established, I think that people just do not answer the phone anymore. The intelligent people who have lives to live either use caller id or answering machines to screen out calls. I refuse to answer surveys because I find that lying to them is more fun.

If the number 17 does not appear on a roulette wheel for 10 hours straight at a casino, they will not care or investigate. Only when the 17 comes up 10 times in a row will they investigate.

When GWB gets 100% of the votes is what it will take for an investigation. At that point the election will be invalidated and GWB will stay on as president just to make sure the investigation is completed and by then we have a long term dicatator. That is what scares me.

[Bill Bored]

16. Breakdown by County Here:

Looks like Lucas Cty. was Diebold which has a thermal printer.

http://www.yourvotecountsohio.org/default.asp?pageLoc=/general/map.html