Anybody who’s not math-challenged want to look at this article and tell me if it’s A Thing? Nathaniel Downes at Addicting Info claims “PROOF: GOP Party Bosses Rigging Elections For Romney“:
The internet has been abuzz since the British website UK Progressive released a report on how a retired NSA analyst Michael Duniho had gone over the Arizona GOP Primary Results from earlier this year and discovered widespread GOP election fraud through what can be called an electronic fingerprint. He found that when you break down the primary results in to their component precincts, and then compare the percentage of each candidate and how much they gained per precinct, a pattern appeared:
__
The larger the district, the larger Mitt Romney’s percentage, in a smooth line.This is statistically improbable to occur even in one election. What he discovered is that this is not a one time fluke at all. Instead, he found the tell-tale signature of electronic manipulation…
As I said: I don’t have the knowledge to evaluate this data, and (as someone who still believes the 1988 election was an electronic trial run) I’m prone to believe the worst of the GOP under any circumstance. So I would genuinely appreciate being told, by people who know what they’re talking about, that this is just one of those anomalous patterns that statistical graphing will throw up.
Real or random, it’s not going to change my conviction that we don’t just need to win this election, we need to win by the widest margins possible.
Maude
I just put the NOAA Sandy update on the last thread.
I’m way inland NJ and I think it could be fierce here.
In the apartment building, there are air conditions in some of the windows and trash cans in the alley.
barath
I was wondering about this voting machine / electoral rigging claim, and I also hope someone with some stats savvy can figure out what’s up with it.
The closeness of the race makes me very nervous – it’s close enough to steal. We need to do everything possible to win this.
The darkness of letting Romney and Ryan, two inveterate liars who care about nothing than a crazy Randian ideology…I just can’t begin to think about it.
Another Halocene Human
As I said on the original post, given that Mitt Romney’s donors live in urbanized areas, why is it at all surprising that larger (read: urbanized) districts broke from Romney? Why wouldn’t his donors be his voters?
And why is it surprising that Sex-Obsessed Grifter Boy Scold’s votes drop the less rural the precinct gets? One assumes people in cities have more internets, the better to know about Sanctimonium’s lovely record in PA.
M. Bouffant
Yup. As a California resident I often vote for a third, fourth or fifth party candidate, sometimes because I’d actually prefer them, sometimes because more votes for a right-wing loon encourages them to get loonier, & scares wussy libs, but not this time.
barath
@Another Halocene Human:
I was wondering if that was it as well, and figure there exists some benign reason as well though it’d take some statistics analysis to figure it out.
But separate from that, having seen electronic voting machines up close and knowing the underlying security of them, I know how vulnerable they are. And like Anne Laurie said, I’m prone to believe the worst as to what they might try to pull to steal the election. That’s where running up the margin matters.
Christian Sieber
This kind of pattern appears in tons of races, not just the 2012 GOP primaries. The authors of the original study found the same patterns grabbing votes for McCain over Obama in the 2008 general election. (I ran the charts myself on Cuyahoga County, Ohio and got the same results.)
Demographically it is hard to explain since you can observe this pattern within a demographically stable county (rather than a more diverse state.) I don’t know enough about the statistics end of that area though.
What I do know is that it is hard to use this as proof of fraud, because fraud on this scale would CLEARLY show up as elections being markedly different from polling in the majority of cases, which seems unlikely for some of the highly scrutinized states where this is claimed to have taken place.
MikeJ
He’s claiming 20 point swings in “stolen” votes. That’s not very believable.
BruinKid
Eh, I’d need more evidence than that analysis. Why do they have the x-axis as a percent? A percent of what? The population size of the largest precinct? Why not just use population numbers? The link also claims Jan Brewer actually lost the Governor’s race. Except all the polls showed she was going to win rather comfortably, even after that horrendous debate performance. So unless they’re also going to claim that the polls were all rigged in the GOP’s favor as well….
Heh, I do like how Ron Paul’s numbers noticeably go down as the population goes up. This’ll most definitely get spread around the Paulbot sites as “proof” of fraud. Or… it could be that in larger populations where people, you know, actually have to live with other people all around them, they realize they actually have to live together in some sort of functioning society, and the whole Lord of the Flies mentality espoused by Paul maybe doesn’t have as much appeal.
Brandon
@MikeJ: This “analysis” is at unskewedpolls.com level of unbelievable. There is simply no basis to afford it any credibility.
Another Halocene Human
@barath: But not all precincts use the same equipment. I see no attempt to correlate election processes with results or some sort of polling/results/turnout/process comparison.
BRADBlog, on the other hand, does look closely at election laws (process), the equipment (and some of it is awful), anomalous results, anomalous turnout, and also suspected or proven cases of election fraud.
Based on the cases highlighted by BRADBlog, I gotta say if there’s fraud in those primary graphs it’s too sneaky to spot right off.
Another Halocene Human
@BruinKid: Heh, I do like how Ron Paul’s numbers noticeably go down as the population goes up. This’ll most definitely get spread around the Paulbot sites as “proof” of fraud. Or… it could be that in larger populations where people, you know, actually have to live with other people all around them, they realize they actually have to live together in some sort of functioning society, and the whole Lord of the Flies mentality espoused by Paul maybe doesn’t have as much appeal.
I just figured because shit like stock markets and insurance firms and departments of economics were concentrated in population centers that there might be more people with a fucking clue about how teh munniez werks, and hence who find Goldbug Paul a bit … out there.
DFH no.6
Prof. Sam Wang at the Princeton Election Consortium (best site for polling analysis, IMHO) says he thinks this report is crazy.
So I’ll go with that.
mdblanche
Man, that is one horribly constructed graph. An analyst who can’t be bothered to properly label what it is they’re graphing does not inspire my confidence.
Xenos
It illustrates and interesting theory of how a very subtle but effective scheme could be done. But this fails the large-scale conspiracy test – doing the exact same type of fraud in different states, across different counties, with different vote counting technologies…. it is just too unlikely to maintain such a complex conspiracy involving so many people.
Not that it is impossible, but just very unlikely. Need some extraordinary evidence for these claims. They can start with a much more complete and properly presented set of graphs and charts.
Xenos
Alternately, this could be a very clever attempt at ratfucking the Paulites and Santorumites in the GOP base. Publishing the claims in a lefty UK site would rather give that game away, though.
It would be nice to see some legit analysis of this issue, eventually.
? Martin
Makes sense, actually. Mitt won the top of the income ladder – and larger districts are the ones that support higher income individuals.
Xenos
A final thought – I am not mathematician, but as a litigator I am used to evaluating evidence and dealing with bullshitters. The moment someone makes claim that is prefaced as “PROOF” it is nearly always bullshit. Scientists and other rigorous thinkers don’t talk like that.
Concerned Citizen
What gets me is, this whole time people have been thinking the media has been pushing the horse race narrative to be self serving. Maybe they have been manufacturing consent like with the Iraq war. If they keep pushing that the race is close nobody, will blink an eye when Romney wins. Add to that pundits trying to discredit accurate pollsters like Nate Silver and well, shit.
Ben Johannson
Perhaps if you express glee at someone’s death, that’ll get the ball rolling. That is what you do, isn’t it?
The Frito Pundito
I downloaded the original data and found that the relation being expressed was between Romney’s percentage of a precinct’s vote and the cumulative total off all precincts with that vote total or lower. Which is why the lines are so straight, and a strange way of summarizing the data. I did a correlation of Romney’s votes in a precinct with the total votes in a precinct (i.e. whether his vote percentage increased with the size of a precinct) and found a r = 0.64 which is likely significant. However, to me this suggests nothing – one can explain it any number of ways, such as Romney spent more time and money in larger precincts, or people in larger precincts tend to be more urban and wealthier, and thus more likely to vote for Romney. This is hardly evidence of vote rigging, and pretty sloppy analysis.
PeakVT
5 AM advisory for Sandy. Landfall is currently projected to be near Atlantic City at around 2AM on Tuesday morning.
TF79
I would agree with Prof Wang and say no, this is not A Thing
Raven
Fucking bedwetters,
SiubhanDuinne
@Ben Johannson:
What??
low-tech cyclist
What the Frito Pundito said. To the extent that larger precincts tend to be in more densely-populated areas, this is just the usual fact that the further towards urban you are along the urban-rural continuum, the more likely you are to be more liberal. And while this correlation has been talked up mostly about Dem v. GOP, it also holds along the more conservative to more moderate/more establishment axis within the GOP.
And like mdblanche said, you gotta wonder about supposedly rigorous analysis from someone who can’t even manage to label their axes in a way that makes sense.
I’m not seeing a smoking gun here. Hell, I’m not even seeing the little whiff of smoke from a cap gun.
Barney
If you go to the original paper, they’ve done this for many 2012 Republican primaries, and also one from 2008 (in which there’s no ‘Romney effect’). My first thought was that Romney, being the ‘reasonable’ candidate among loons, would do better in the more populated areas where there’s less, shall we say, ‘eccentricity’ (and that might tally with Gingrich being level – not as weird as Paul or Santorum, just dishonest). But the interesting graph, from that point of view, is New Hampshire – the one state where Huntsman, more reasonable than Romney, was competitive. But the “Romney wins big precincts’ effect is even more pronounced there, while Huntsman’s line gradually slopes down (not as fast as Paul’s does, though).
All I’ve got left to suggest is that the Romney campaign targeted the high population precincts well, and so got the most bang for their buck. There’s nothing in the report about which precincts count by hand, and which by machine – BradBlog said 40% of ‘towns’ count ballots by hand, but that only accounts for 10% of total ballots. I’d like to see some comparisons of hand-counted and machine-tallied ballots.
PopeRatzo
@Christian Sieber:
Unless the polling is bogus too.
Christian Sieber
Yeah, but think about what that would entail, especially in a hotly contested and highly scrutinized race like Ohio 2008 General or Florida 2012 GOP-Primary — both of which had this “effect” show up in pronounced ways, but there would have been an intense hue and cry had there been intense discrepancies from the expected results. In major races it would be extremely difficult for the polling to be bogus in a way that would make this grade of election-stealing not noticeable. Not every outfit is Rasmussen, after all.
Schlemizel
@BruinKid:
My boss is a paultard – I am going to make sure he sees this & blames the GOP for their failure to nominate the one true savior.
If we can inflame the randoids it might separate them from their support for the GOP. Any chuck that could be taken out of the GOP block, if it votes 3rd or not at all, would be a good thing
max
Well, this is something to wake up to when bleary-eyed. So I too downloaded the Excel file and what Frito Pundit says: the ‘pattern’ is based on simply adding the cumulative vote totals together starting with the smallest precincts first, which means Gingrich starts off with 100% of the total vote based on the one-vote precinct he won.
What the guy has demonstrated is the Law of Large Numbers, basically. As the vote total grows the final percentage outcome gravitates toward the final total. Woo. (Extra complication: this appears to Pima county, so precincts populated by lots of R’s (because they’re rich? because they’re Mormon? Both?) show lots of R votes, and Romney is popular among… R’s.
To show machine generated fraud (which does not have to employ the algo this guy thinks is the culprit) you’d need to compare any given precinct non-cumulative total to the actual outcome. That still wouldn’t tell you anything, exactly, since that just could be tracking shifts in demos.
Whatever, this would need to be a lot more sophisticated and include more than just one county before we would clear the plausibility hurdle.
That said, there are excellent reasons to get Congress to require that electronic voting machines used in Federal elections (they can do that!) generate printed copies of each voters ballot choices. (To prevent voters from skipping with the ballots, they can just make ballots visible through plexiglass so the voter can confirm the ballot without being able to touch it.)
max
[‘Too early in the morning for this.’]
28 Percent
Chart is bogus, and basically the only safeguard against tampering on evoting machines is the honesty and diligence of the system auditors, which should scare the ever lovin’ shit out of you.
28 Percent
Chart is bogus, and basically the only safeguard against tampering on evoting machines is the honesty and diligence of the system auditors, which should scare the ever lovin’ shit out of you.
Poopyman
All this while I slept? Thanks to all of you! I did a once-over of the linked article and it stunk like a 5-day old fish. Graphs that don’t really say anything but just trend downward for non-Mitts, strident claims of “proof…. Pretty boilerplate tinfoil hat article, to tell the truth. Glad you folks actually ran the numbers.
Poopyman
Oh, BTW. Don’t look now, but DKos is pushing this story.
Omnes Omnibus
I am getting really tired of this shit. The Bilderbergers staged the moon landings in Area 51; they arranged it at the Bohemian Grove in order to let the Illuminati have control of the the US gold supply. Look what happened to Bretton Woods! Wake up, sheeple!
Vote, GOTfuckingV, and look out for voter suppression. Voter suppression efforts are real; fight against something that exists rather than looking for shadowy conspiracies. Fuck.
MaxL
@The Frito Pundito I think you found the exact same error as this commenter on the original UK Progressive article:
“As much as I’d like to believe this, the analysis of the data is incorrect. The cumulative value is used incorrectly on the X-axis, instead of just the Y, giving statistical bias to the large numbered precincts due to the accumulation. If you use column L instead, which is the integer total number of votes per precinct, you get the flat line you are expecting with the cumulative still used on the Y axis. If you generate the resulting graph it also produces a lot of variance in the smaller precincts before it levels out, which is what I would expect intuitively, it’s unlikely that smaller precincts will have such a consistent curve shown in your graph with such a small amount of data, ie 10 voters. I can post the picture of your modified graph, but this comment section does not allow it. If you send me your email, I can send it to you along with the modified excel sheet.”
Shinobi
He lost me as soon as he argued that you’d expect not to see significant correlations in the between precinct data. When dealing with things like votes, or demographics, nothing is truly independent.
After that I don’t even need to look at the data to know this guy is most likely talking out of his ass.
(But high fives to the people who did. It is too f-ing early to work.)
gorillagogo
@Another Halocene Human:
From my understanding, it’s not the fact that Romney received more votes from urban areas, it’s that his percentage of votes increased in a linear fashion as the district size increased. That is about as likely to occur as winning the lottery and getting struck by lightning on the same day.
The fact that this same pattern emerges over multiple elections, always favoring Republican candidates but only in states that are controlled by Republicans tells us that the fix is in.
dr. bloor
@Poopyman: Oh hell, it’s a diarist. It’s like one of us posting it in the comments and claiming Cole is pushing it.
Matt McIrvin
@DFH no.6: Yeah, but Sam Wang is citing me, so, y’know, I wouldn’t call his sources the most authoritative.
merrinc
I can’t stand Addicting Info. They’re not quite the Fox News of the left but they’re close, IMO.
karen marie
People have been fighting in Pima County, AZ since 2006 over alleged election fraud that year. The election board and the county both concede that “cheating is easy” but refuse to open the books. I haven’t gotten down into the nitty-gritty.
Higgs Boson's Mate
‘Scuse my mathematical ignorance, but it looks like the author just proved that there are people who consistently vote Republican in Republican-controlled states.
Gian
while this particular wannabe expose isn’t all that credible, we’ve seen that county clerk in WI mess with the integrety of the vote.
http://expressmilwaukee.com/article-14489-incompetence-or-fraud-.html
I don’t have any idea how much of the GOP actions are involved in messing with actual ballots, versus just making it hard to get the chance to vote.
and I don’t mean just the crop of new ID laws this year.
I mean the GOP operative convicted in NH of jamming the Democratic party’s phones so they couldn’t make the get out the vote calls
I mean the 2004 Ohio action where if you voted in a majority Democratic area, you were screwed with the ratio of voting machines to voters. How often do you see stories about rich suburban republicans having to wait 8 hours in the rain to vote?
at sopme level they feel it’s their birthright to rule, and screw democracy when it gets in the way… they’re happy to remind us that it’s a republic not a democracy… on the fox-approved talking point
Tim I
This CT bullshit is getting far too much play in progressive blogs. It is utter garbage with no statistical support. Just a couple of guys in tinfoil hats having a circle jerk.
J R in WV
I was a software engineer/project manager for nearly 30 years.
Bank ATMs keep a paper record of every transaction and balance to the penny, as well as (often) a video record of who walked up to the control panel.
But the same manufacturers that build ATMs also build voting machines without a running paper trail. Here in WV my rural county (which has had a lot of experience with vote-rigging) uses computerized touch-screen voting machines – and every one keeps a paper trail of every voter decision, which you can read through a glass window.
Any computer can be owned by someone with “root access” to the machine, and often without special security access, just physical access and time.
I think we should be using paper ballots, with big number two pencils to mark who we want to vote for. It works for other countries, and you can recount them easily. There is no easy way to cheat with physical ballots, if reps from political parties are with the ballot boxes when they’re sealed and when they’re opened.
Computers that you can get your hands on are way too easy to “adjust”. This analysis seems to me to be a clear warning that there is a nation-wide conspiracy to change vote counts with certain brands (if not all brands) of voting machines.
In my mind, this is a treasonous conspiracy against the foundation of our country and way of life!
Rand Careaga
(Time for a little meme-whoring here)…
The classics never go out of style: The Diebold Variations.
MikeJake
It doesn’t really surprise me that support for Ron Paul and Rick Santorum declined as you move out of the boonies.
Omnes Omnibus
@Gian:
No, Kathy Nickolaus fucked up reporting; she did not mess with votes. Here is a link with the reports of both the WI GAB and the independent investigation into irregularities in Waukesha County during the Supreme Court election.
Fundamentally – stupid, not evil.
Another Halocene Human
@gorillagogo: From my understanding, it’s not the fact that Romney received more votes from urban areas, it’s that his percentage of votes increased in a linear fashion as the district size increased.
Yes, dude, and I’m arguing that his percentage of votes increasing in urban areas is exactly what we’d expect to see.
Next.
Just Some Fuckhead
Wouldn’t the proof come from examining the trend lines in electronic voting districts v. manual ones?
JackHughes
Not nearly as damning as the results of the 2010 Brown-Coakley race in Massachusetts, where Brown out-polled Coakley in more liberal urban areas using paperless DREs, and Coakley out-polled Brown in more conservative rural areas using hand counts.
Those results just ain’t right.
Gian
@Omnes Omnibus:
Gotcha, or got me – I cop to having a bias in that I just have a hard time trusting the GOP to be honest about anything. (I mean this is a party with a central article of faith in the laffer curve preached to suckers to sell tax cuts for the rich)
wikipedia has a page on the NH scandal:
http://en.wikipedia.org/wiki/2002_New_Hampshire_Senate_election_phone_jamming_scandal
The 2002 New Hampshire Senate election phone jamming scandal involves the use of a telemarketing firm hired by that state’s Republican Party (NHGOP) for election tampering. The tampering involved using a call center to jam the phone lines of a Get Out the Vote (GOTV) operation. In the end, 900 calls were made for 45 minutes of disruption to the Democratic-leaning call centers.
During that state’s 2002 election for the U.S. Senate seat being vacated by Robert C. Smith, the NHGOP hired GOP Marketplace, based in Northern Virginia, to jam another phone bank being used by the state Democratic Party and the firefighters’ union for efforts to turn out voters on behalf of then-governor Jeanne Shaheen on Election Day. John E. Sununu, the Republican candidate, won a narrow victory. In addition to criminal prosecutions, disclosures in the case have come from a civil suit filed by the state’s Democratic Party against the state’s Republican Party (now settled).
Four men have been convicted of, or pled guilty to, federal crimes and sentenced to prison for their involvement as of 2008. One conviction has been reversed by an appeals court, a decision prosecutors are appealing. James Tobin, freed on appeal, was later indicted on charges of lying to the FBI during the original investigation.
Jebediah
@max:
I wouldn’t expect such a requirement to make it through a Congress with any significant Republican power, but if it were introduced it might be instructive to see where the loudest squealing comes from.
RSA
I haven’t checked the data, but when I read a phrase like this, I expect an explanation that actually gives me the probability of seeing the pattern presented. That doesn’t seem to be part of the article.
Omnes Omnibus
@Gian: As I noted above, voter suppression is real and worrisome. I think that is where people should focus their efforts. Even if one worries about voting machines, they can only affect things at the margins. Working on GOTV and making sure that things aren’t close enough to steal takes machine tampering out of the picture.
Rob. S
Very credible researcher, Google this: “NSA analyst Michael Duniho.”
NSA stands for National Security Agency, by the way.
It is not a tight race, how can it be? All these polls? Fiction.
Spatula
Not to worry…the Democrats will continue to yell and scream for change and safeguards, and pass laws to prevent this kind of thing, and shout from the rooftoops on an endless loop about how Republicans are stealing elections, just as they have done since the coup of 2000.
Oh, wait…
Roger Moore
@Jebediah:
Republican voting officials complaining about the cost of printers for all the electronic voting machines. “You made us spend all this money on new voting machines, and now you’re making us go back and spend more!” We know this because it’s already happened any time the issue is raised.
Steve Finlay
@The Frito Pundito: Thanks very much for checking. I could not figure out what the x axis was supposed to mean. Yes, setting up the x axis as cumulative is bizarre. It’s also likely to create the appearance of a pattern when there is none.
Joey Giraud
@Xenos:
How do you know if your model of “the conspiracy” is the only possible one?
And might not some “conspiracies” be more loosely co-ordinated, perhaps even “crowd-sourced?”
I’ve always found this “it’s too complex, someone will spill the beans” argument to be facile, as if all conspiracies must be like Mission Impossible.
And the idea that even a single whistle blower could topple the conspiracy as if it were a house of cards. Looking at what happens to real world whistle blowers should teach us that things don’t work that way.
Rob. S
“And the idea that even a single whistle blower could topple the conspiracy as if it were a house of cards. Looking at what happens to real world whistle blowers should teach us that things don’t work that way.”
The founders where whistle blowers. In a democratic sense, they eradicated the Despotism. Minus the votes, where is democracy? And they speak of Assad? How hypocritical.
joel hanes
@max:
As a computer engineer, it’s my opinion that there are excellent reasons to completely prohibit the use of computers as primary and secondary tabulators in the election process, and to insist on tangible primary ballots that can be audited.
owlbear1
@Poopyman:
I wouldn’t call 4 recommendations and 11 comments(ridiculous being the most prominent response) “pushing” but hey if you see a DKos conspiracy, go for it.
Joey Giraud
@Rob. S:
Weird response. Makes no sense.
Rob. S
@Joey Giraud:
Read the document: “NSA analyst Michael Duniho.”
Rob. S
@Joey Giraud:
Another academic source “Francois Choquette, James Johnson”
Not at all weird makes plenty a sense.
Bob In Portland
When I first heard about Bush’s “Help America Vote Act” I thought, why the hell do Republicans want to help people to vote? I figure it’s part of the business plan.
My favorite outcome is that Obama wins outright but Tagg is indicted.
kwAwk
I know I’m a bit late to the party on this one but I also downloaded the data and re-arranged it. Instead of looking at the cumulative affect of the data I looked at the data by individual precinct.
I created four different bar graphs showing each of the four candidates by precinct from smallest to largest and what I came up with is that for Gingrich and Santorum their support remains rather steady across the spectrum.
Where Romney’s gains came from in the spectrum from smallest precinct to largest was mostly from Ron Paul. Drawing a trend line on each of their charts would result in lines that have almost the opposite slope.
This would seem to make sense if you consider Paul’s anti-government message and Romney’s pro-business message to be two sides of the same coin.
It also makes sense in the Paul’s libtard voters wouldn’t find Santorum’s big government social conservatism appealing nor would they find Gingrich’s pragmatic work within the system type of conservatism appealing.
That’s my two cents at least.
Joey Giraud
@Rob. S:
I’m sure you make sense to yourself, but links to the OP story don’t lend a meaning to your word salad.
Unless you mean that Duniho is a whistle blower but can’t seem to topple that house of cards.
But statistical “proof” isn’t proof at all; it’s statistics.
The only way to prove voting machine fraud is to actually count real votes and compare the numbers to what the machine says.
Proof must be comprehensible, few people understand statistics.
Everyone understands: Five hundred votes for Obama, the machine said four hundred. FAIL.