There's a problem with electronic voting machines in the USA
August 7, 2015 7:33 AM   Subscribe

My statistical analysis shows patterns indicative of vote manipulation in machines. The manipulation is relatively small, compared with the inherent variability of election results, but it is consistent. [...W]e have a serious pervasive and systematic problem with electronic voting machines.

Her summary: The only way to prove vote fraud is through a post-election audit demonstrating significant deviations from the reported totals. That is what I want to see done. It is also recommended by experts in voting integrity, including 'The American Voting Experience: Report and Recommendations of the Presidential Commission on Election Administration' published in 2014.

However, gaining access to more data has been problematic.

The research was done by Beth Clarkson, who has documented her progress (you need to follow "older posts").
posted by andrewcooke (74 comments total) 29 users marked this as a favorite
 
This is my surprised face.

(Note: I'm not actually surprised)
posted by caution live frogs at 7:51 AM on August 7, 2015 [7 favorites]


I guess this is like the difference between discovering a software bug and having a working security exploit. It was obvious back during the Bush II administration that doing this was trivial, but it's good to have concrete evidence that this is happening. Maybe this will finally lay to rest the asinine idea of electronic vote casting.

I mean, it probably won't, but it's important to have hope, right?
posted by indubitable at 7:57 AM on August 7, 2015 [2 favorites]


The SoS, Kris Kobach, is a racist nutbag (which is quite an achievement for Kansas) who has made it a personal crusade to disenfranchise non-white voters and encourage voter intimidation. Blocking access to evidence of electoral fraud while banging on about essentially non-existent voter fraud is just par for the course.
posted by zombieflanders at 7:59 AM on August 7, 2015 [11 favorites]


Most of these show a slight linear trend upward. That, to me, is evidence that Republicans vote later in the day.
posted by MisantropicPainforest at 8:06 AM on August 7, 2015


OK, not that black-box voting isn't a worrying issue, but I don't buy this analysis one iota. I remember the exact same methodology, with exactly the same type of graphs, being used as "evidence" that Ron Paul got robbed in the 2012 election. (e.g. in this extremely long Reddit thread) The problem, both now and back then, is that the null hypothesis:
If there is no significant interaction between voting precinct size and %Republican (%R) vote (our null hypothesis), then we expect no significant slope to the ratio of the cumulative sums of Republican votes to total votes cast for sufficiently large n.
is an extremely strong assumption. Finding that it's violated isn't really evidence of anything in particular; there are all kinds of demographic effects that would cause voting preferences to be different between large and small precincts. So coming up with p-values based on that hypothesis, and then just casually asserting (with no justification) that electoral fraud is the most likely explanation, is pretty disingenuous.
For the concern regarding voting machine manipulation, I restricted the visual inspection to the right half of the graph showing the larger precincts associated with higher density population centers. This truncation of the data eliminates the confounding effect of rural districts which are expected to have both a higher %R vote and smaller precincts in terms of number of votes cast.
(emphasis mine) No it doesn't; even if you only look at part of the graph, it's based on a cumulative sum of all of the precincts. So pretty much any statistical difference between large and small districts will show up as a linear trend that looks exactly like this.

The article ignores an obvious question: assuming someone was manipulating the electronic voting machines, why in the world would they program it to rig the votes by a different ratio in precincts of different sizes?
posted by teraflop at 8:07 AM on August 7, 2015 [34 favorites]


We get upset at some people when they don't seem skeptical enough (voting machines, Internet privacy issues, big agriculture practices), yet we also get upset when people do question technology (related to climate change, vaccinations, etc.).

I get that these are frequently different sets of people, but it may not seem so to any individual. Also, politicians do come from and are influenced by the people who elect them (thank goodness).
posted by amtho at 8:07 AM on August 7, 2015


why in the world would they program it to rig the votes by a different ratio in precincts of different sizes?

In order to further obfuscate the fact of the rigging?
posted by Rat Spatula at 8:11 AM on August 7, 2015


There have been indications that Diebold (whose former CEO, Wally O'Dell, was a high-profile fundraiser for George W. Bush) and Election Systems & Software voting machines have been used to rig elections as far back as 2002. The security on the Diebold machines is trivially weak and has remained so for over a decade.
posted by ryanshepard at 8:15 AM on August 7, 2015 [10 favorites]


The differing ratios could be an artifact of the software exploit that was used, maybe.
posted by Vulgar Euphemism at 8:19 AM on August 7, 2015


We get upset at some people when they don't seem skeptical enough (voting machines, Internet privacy issues, big agriculture practices), yet we also get upset when people do question technology (related to climate change, vaccinations, etc.).

I wouldn't have thought it possible to so readily conflate skepticism and denialism, because they derive from opposite approaches and mindsets. You might add evolution to your latter list. None of them are skepticism because they begin with a belief, arrived at by choice, and won't accept any information that contradicts it. Skepticism is about rejecting belief in favor of information. Denialism is about rejecting information in favor of belief.
posted by George_Spiggott at 8:20 AM on August 7, 2015 [33 favorites]



The article ignores an obvious question: assuming someone was manipulating the electronic voting machines, why in the world would they program it to rig the votes by a different ratio in precincts of different sizes?


Because in a small rural precinct, you could literally provoke all the voters to return, assemble in a small gym, and demand to know why their vote totals don't match what they say they voted for.
posted by ocschwar at 8:20 AM on August 7, 2015 [3 favorites]


why in the world would they program it to rig the votes by a different ratio in precincts of different sizes?

As Rat Spatula said - in order to hide the process.

Were I programming the machines to be biased, I'd start by always recording the first hundred or so votes correctly and some reasonable number of votes to the side I did not favor - then include increasing amounts of bias, randomly flipping votes to the side I was being paid to support at some slightly increasing(but capped) percentage - so it might start at 10% then slowly increase to 20% as the number of votes increased. I'd also flip a few votes the other way to maintain a level of plausible deniability (at least until someone looked at the source code - but that has been traditionally been kept a trade secret).

So precincts of different sizes would naturally have different ratios of rigged votes.
posted by Death and Gravity at 8:21 AM on August 7, 2015 [6 favorites]


I would love to do the same analysis for the Georgia 2002 election, whose results were sketchy as fuck. 13 years later, we're still using the same Diebold machines and there still is no paper trail.
posted by hydropsyche at 8:30 AM on August 7, 2015


I'm not saying that voting manipulation doesn't occur, but to prove it, you'd have to show that unexplained variances are related to a variable uncorrelated with partisan vote percentages. The problem is that population size and partisan vote percentages are highly correlated, although there are some exceptions (rural blacks in the Southern Black Belt, rural Hispanos in New Mexico).
posted by jonp72 at 8:33 AM on August 7, 2015 [1 favorite]


Hand counted ballots seem to work pretty well in Canada.
posted by Evstar at 8:35 AM on August 7, 2015 [15 favorites]


you'd have to show that unexplained variances are related to a variable uncorrelated with partisan vote percentages.

Not necessarily, at least with this analysis, since she is examining changes over time, so what you would need to control for would be anything that explains the changes in vote share over time.
posted by MisantropicPainforest at 8:39 AM on August 7, 2015


My reading of the article is that:

1) in precincts without voting machines, there is no relationship between number of votes and %R

2) in precincts WITH voting machines, there is a statistically significant relationship between number of votes and %R

I'm skeptical. Is there some way to explain this discrepancy that doesn't involve fraud?
posted by Avenger at 8:39 AM on August 7, 2015 [3 favorites]


Is there some way to explain this discrepancy that doesn't involve fraud?

If the treatment assignment (which precints get voting machines) is correlated with anything that is not fraud that would affect the percentage change over time, then that would explain it without reference to fraud.
posted by MisantropicPainforest at 8:42 AM on August 7, 2015 [5 favorites]


I don't really buy the analysis either, but that's sort of beside the point. The problem with the machines is that you can't, by design prove that they are working correctly.

They are always going to be open to accusations of bias and election-rigging with these systems, because they're black boxes, the internal workings of which aren't open for examination (and are non-trivial to examine even if they were -- I personally don't even think that open source electronic voting would be sufficient, because the number of people capable of auditing the source code is just too low; the only good solution is one that's transparent to the average voter, which means paper). So even though I don't really find the analysis to be a smoking-gun demonstrating that an election was rigged, I don't think that there's enough evidence to say conclusively that isn't what happened either, and that is really, really troubling.

If there's any chance that your vote isn't going to count, that's corrosive to the whole idea of democracy. And I tend to wonder if that isn't why the machines remain around: you don't actually need to rig them in order for them to have an effect. Even if they work perfectly in terms of counting votes, if you keep them around, keep their inner workings a secret, and spread the rumor that they result in rigged elections ... well, what better way to keep people at home than to tell them that their vote won't count or influence the results anyway? And if by staying home they make that perception into reality...? Game, set, match.
posted by Kadin2048 at 8:44 AM on August 7, 2015 [22 favorites]


Is there some way to explain this discrepancy that doesn't involve fraud?

If the treatment assignment (which precints get voting machines) is correlated with anything that is not fraud that would affect the percentage change over time, then that would explain it without reference to fraud.
posted by MisantropicPainforest at 8:42 AM on August 7 [1 favorite −] [!]


Yeah that's a good point -- the assignment of voting machines throughout a county or state is almost certainly not random ... so an actual study would have to control for this somehow.

Thanks for refreshing my Research Methods knowledge!
posted by Avenger at 8:45 AM on August 7, 2015 [1 favorite]


I'm skeptical. Is there some way to explain this discrepancy that doesn't involve fraud?

Relative wealth of the districts? I've never seen data for Ohio, but the UK voter maps for cities generally have the wealthier areas leaning right. How expensive are voting machines?
posted by metaBugs at 8:45 AM on August 7, 2015 [2 favorites]


See, this is the exact heart of the the problem:

> I don't really buy the analysis either, but that's sort of beside the point. The problem with the machines is that you can't, by design prove that they are working correctly.

There may not be fraud. There may be. But there's no excuse - no excuse - for our inability to prove it, one way or the other. It is the simplest thing in the world to keep a paper trail. It is very simple to just separate the two steps: use one machine to prepare your vote, and use a different machine to count it. In between, you can verify what was recorded, and we're done. Being told "Just trust us" by an unaccountable software company is the bit that's unacceptable.

> We get upset at some people when they don't seem skeptical enough (voting machines, Internet privacy issues, big agriculture practices), yet we also get upset when people do question technology (related to climate change, vaccinations, etc.).

Ok, I'm guilty on both counts. But I really don't see the discrepancy. If there is objectively verifiable information, and people without obvious conflicts of interest do verify it and claim there is a problem, then we should take it seriously. If there isn't, we shouldn't.
posted by RedOrGreen at 8:59 AM on August 7, 2015 [8 favorites]


Oh this is old news. Verifiable reports of Diebold voting machine fraud and related lack of security have been underreported for almost a decade now. In fact a major I.T. figure responsible for those machines and the Bush election fiasco, Mike Connell, was being subpoenaed to testify in a voter fraud lawsuit but he conveniently died in a light plane crash before he could testify.
posted by AGameOfMoans at 9:07 AM on August 7, 2015 [10 favorites]


It might be somewhat defensible if electronic voting were actually solving some kind of problem, but it's not. It's purely a case of a solution looking for a problem. There is no reason to entertain them, at all. They should be tossed out completely.
posted by odinsdream at 9:11 AM on August 7, 2015 [5 favorites]


The problem is twofold: (1) Republican donors need government contracts to profit and (2) black people are voting
posted by T.D. Strange at 9:13 AM on August 7, 2015 [1 favorite]


1. US voting machine systems are so leaky from end to end that it's inconceivable that there isn't a huge amount of voter machine fraud.

2. But this doesn't constitute proof.

3. The statistical anomalies in the linked article are suspicious and if I were an honest person running the voting system, I'd be worried by them - but there are all sorts of reasonably plausible mechanisms to explain them without fraud.

4. So they don't constitute proof.

Point is, the system has been set up so it's impossible to verify, so there's no encryption used (strong or weak), so there are no checksums, that mechanical safeguards are trivial to bypass. Heck, in many cases the data is set up as Excel spreadsheet.

As such, it wouldn't take a genius to "undetectably" forge votes. I'd expect to be able to write and such a program in a day - if it were to run on my own hardware (but in many cases the votes are stored on a removable card that I could just take out, put in my laptop, run the program, replace).

So the chances of finding a smoking gun should be very small. Heck, I'm surprised this leaked out, if it even did... you wouldn't have to be a statistician to come up with a fake that was impossible to tell from the real thing (though you'd need "mathematical maturity" I suppose).
posted by lupus_yonderboy at 9:15 AM on August 7, 2015 [2 favorites]


I think folks are going at this from the wrong angle. It's not that people aren't committing fraud or are committing fraud, we don't know. We might have the capacity to know, but much like NSA surveillance or programs to undermine encryption, you don't have very good luck determining with a degree of certainty that satisfies doubters or that particular breed of pedantic internet Poindexter we're all familiar with.

If you want to do something about electronic voting security, just start an electronic security consulting firm, like The Hacking Team or other exploit vendors, and target electronic voting specifically. Sell voting machine exploits to governments, so you can claim you're still ethical, and I think you'd see a lot more movement on this front.
posted by turntraitor at 9:20 AM on August 7, 2015 [1 favorite]


> If you want to do something about electronic voting security, just start an electronic security consulting firm, like The Hacking Team or other exploit vendors, and target electronic voting specifically.

You'd go to jail so fast your head would spin. There are all sorts of felonies with long jail sentences involved. They could easily throw on "conspiracy" charges too.

> Sell voting machine exploits to governments,

What?! Why would they pay you money? They are the ones who set up these broken systems!
posted by lupus_yonderboy at 9:25 AM on August 7, 2015


Hand counted ballots seem to work pretty well in Canada.

You might be astonished. A typical ballot in the U.S. might have 90 or more different votes on it: city officials, county officials, state officials, federal officials. At each level there might be judges, prosecuting attorneys, land commissioners, water commissioners, insurance commissioners, highway commissioners, sheriffs. And in addition a dozen or so ballot initiatives for everything from tax bonds to dog leash laws.

It's a nutty system in which most people just guess at a bunch of votes they know nothing about. And if counted by hand, it might take weeks to get results.
posted by JackFlash at 9:26 AM on August 7, 2015 [4 favorites]


Also, there’s the fact that the polls are red-shifted (where there is systematic biasing toward Republican candidates) and have been for several election cycles. This is routinely assumed to be due to Republicans being less likely to answer pollsters, but there is no empirical justification for it. It’s used by polling organizations in their models in order to more accurately predict official results.
In the UK we call this the "Shy Tory" phenomenon, which I feels covers the situation more appropriately: right-wing voters ought to be ashamed of themselves; some are at least smart enough to act ashamed.
posted by rum-soaked space hobo at 9:30 AM on August 7, 2015 [1 favorite]


It's a nutty system in which most people just guess at a bunch of votes they know nothing about. And if counted by hand, it might take weeks to get results.

And here's the problem. We are so impatient about this that we need immediate results; no, we need results before the voting starts, with polls, and polls-of-polls. It's completely insane.

It is in fact entirely reasonable to count votes by hand and wait a few days, even weeks, for the results. Elections aren't held the day before winners start their jobs. If that's what it takes to have honest representation, damn straight it should be done by hand.
posted by odinsdream at 9:34 AM on August 7, 2015 [17 favorites]


Why would it matter if results weren't available on election day?

Elections in Washington state (the system I'm most familiar with) are entirely vote-by-mail; as a result of this, typically it takes about a week for the wave of votes that are posted at the last possible minute to come in and get counted. As such, in close races the outcome isn't known for sure until a good while after election day. This has produced, as far as I can tell, absolutely no problems for anyone.
posted by You Can't Tip a Buick at 9:35 AM on August 7, 2015 [7 favorites]


And if counted by hand, it might take weeks to get results.

So? It's not like Westminster systems where the new term begins basically the next day. Instant results aren't necessary except perhaps at the Presidential level so transition teams can get to work. That can be counted quickly by hand, and then take time over the next days/weeks to count up the rest.
posted by feckless fecal fear mongering at 9:35 AM on August 7, 2015 [1 favorite]


It might be somewhat defensible if electronic voting were actually solving some kind of problem, but it's not.

That's not quite true, though the problems solved are smallish. Electronic voting machines mean that you can't run out of ballots in large print or in Spanish/Chinese/Bengali/Korean/etc. Electronic voting machines also make it impossible to invalidate a ballot with an overvote, and could (though I don't think any do right now) alert the voter to undervotes (you didn't vote for US House; was that your intent?).

It's a nutty system in which most people just guess at a bunch of votes they know nothing about. And if counted by hand, it might take weeks to get results.

It's indeed bonkers, but the problem isn't that it takes a while. The problem is that hand-counting a ballot with 50-100 individual elections on it by hand is virtually certain to be more error-prone than using a machine to do it. For the same reason that hand-grading a 50-100 item scantron test is less accurate than running it through the machine.
posted by ROU_Xenophobe at 9:45 AM on August 7, 2015 [4 favorites]


It is in fact entirely reasonable to count votes by hand and wait a few days, even weeks, for the results.

It's not just the time. It's also the expense. And the manual errors. And the difficulty of maintaining security for all of that time with hundreds or thousands of ballot counters coming and going for weeks.

Electronic voting is much more practical, with paper ballots as a backup for recount if needed. You vote electronically, the machine records the vote and prints out a paper ballot that the voter can quickly verify and the paper ballot is saved for backup. Only in a very close election would anyone ever need to count paper ballots.
posted by JackFlash at 9:54 AM on August 7, 2015 [2 favorites]


Electronic voting, as a concept, is a good idea. You reduce the amount of errors for both the person voting (the ballot is checked for errors before submission), and the recorder (no human error from tallying up votes). The problem is that we have black box voting machines controlled by private companies. A federally funded program to construct a voting machine and have the specs and code out in the open, followed by funding to run cracking teams at prototype machines would make something more secure than what we have now. If you print out paper ballots for later auditing (and do random spot audits), we would have a great system.
posted by demiurge at 9:57 AM on August 7, 2015 [4 favorites]


You vote electronically, the machine records the vote and prints out a paper ballot that the voter can quickly verify and the paper ballot is saved for backup.

Since you wouldn't carry your paper ballot out of the polling place anyway (concerns about vote-selling), then the purpose of the electronic voting machine should be to produce the paper ballot and nothing more. You then read it to verify it's correct, and stick it in the machine near the door which scans it, and that's where your vote gets tallied.

This way the individual voting machines wouldn't need to be audited, removing a point of failure from the system and reducing the difficulty and expense of an audit. Only the scanners and tabulators would need to be auditable; your voters would be verify the voting machines as they used them.
posted by George_Spiggott at 9:59 AM on August 7, 2015 [2 favorites]


The thing is Diebold has proven for decades that they can build secure, error-free machines--check out the logo on the next few ATMs you use.
posted by feckless fecal fear mongering at 10:01 AM on August 7, 2015


Another virtue of electronic voting machines is that the order the candidates are listed can be randomized for each voter to mitigate the issue of the higher the candidate is on the list, the more votes they get. Of course, this doesn't solve the issue of uninformed voters who don't know who the candidates are just choosing someone who's more likely to be near the top of the list.
posted by ShooBoo at 10:06 AM on August 7, 2015


A 2013 federal indictment (which they settled later that year; apparently "settling" an indictment is something you can do if you're a corporation), appears to have prompted this history of Diebold in Ohio alone. Read it and weep.

(What's worse is that's just Diebold. The shenanigans in Ohio went far beyond any one company.)
posted by George_Spiggott at 10:07 AM on August 7, 2015 [1 favorite]


I personally don't even think that open source electronic voting would be sufficient, because the number of people capable of auditing the source code is just too low
I personally am certain that open source electronic voting would not be sufficient, even with an infinite number of Underhanded C Contest winners auditing the source code, because once you turn a voting machine into something close enough to a Universal Turing Machine, you run into the problem that for every program X of the form "count votes correctly, and when my code is requested return a copy" there exists another program Y(X) of the form "count votes incorrectly except during audits, and when my code is requested return a copy of X instead" which can be loaded onto the same hardware.

Open source electronic-voting-to-produce-printed-ballots is probably as good as we're going to get, and even then we'll have the constant temptation to do the counting electronically and ignore the audit trail.
posted by roystgnr at 10:08 AM on August 7, 2015 [6 favorites]


lupus_yonderboy, you are betraying the spirit of your name!

You'd go to jail so fast your head would spin. There are all sorts of felonies with long jail sentences involved. They could easily throw on "conspiracy" charges too.

Why would this be any more illegal than the type of exploits these security research firms already sell to governments? You can buy exploits to break into people's phones, to impersonate them for authenticated systems, even to destroy computer-controlled equipment, all of which are felonies with long jail sentences involved.

I'm just saying: let's expand the range of devices that are considered attack surfaces.

As for why a government, being ostensibly in charge of these systems, would purchase exploits for them? Well, two possible reasons:

1) Naively, one could imagine they'd want to better secure the systems and enhance the credibility of the franchise.

2) Deniability. Just one party operative with a USB stick is much easier and less traceable.
posted by turntraitor at 10:19 AM on August 7, 2015


It's been said above, but again, the only alternative to unprovable black box voting is not hand counting of ridiculously long ballots. All you need is separation of vote preparation from vote counting.

Use whatever machine you want to prepare a ballot - complete accessibility, all possible languages, never run out of ballots. Then have the machine spit out a physical, verifiable object - the simplest would be a filled scantron paper ballot, but use punch cards, or whatever floats your (or your campaign donor's) boat.

Then, use a different machine to count the votes, while storing the physical tokens. And randomly audit 0.1% of precincts nationwide. That's it.
posted by RedOrGreen at 10:21 AM on August 7, 2015 [4 favorites]


Let's be clear here: the only two options are not "hand counting by humans" and "electronic voting with no paper trail". My precincts have always used a version of scantron where you fill out a ballot and then it goes into a machine for counting. The actual ballots remain available for recounts as needed. You could do the same with an electronic voting machine: have it print out a paper copy that is reviewed by the voter before putting into a secure box, that is later available for audits.

Voter fraud is not easy to pull off if you have paper backups of how people voted.
posted by zug at 10:25 AM on August 7, 2015 [2 favorites]


The HBO documentary Hacking Democracy [1h20m] came out nearly 10 years ago and demonstrated not only that there are sufficient doubts about vote totals from electronic voting machines to warrant serious questioning using them, it also demonstrated for the camera that the Diebold machines can be hacked pretty easily to allow vote totals to end up however the hackers want them to end up.

Whether there is actual bias baked into the software or not, the black box nature of the devices should be enough to make all precinct captains want to revert to the old "walk behind the curtain, flip the levers, and pull the handle to register the votes and open the curtain" manual voting machines.

(Although WA and OR have vote-by-mail which has always felt like the best option to me. Here's a ballot weeks before election day! Here's a giant voter information book too! Fill it out at your leisure and get it back to us (either by mail or by secure drop-off location) before election day! No pressure of remembering exactly which candidates or what side of various proposals you wanted to choose!)
posted by hippybear at 10:38 AM on August 7, 2015 [1 favorite]


At some point, the cries of voter machine fraud start to resemble every other case of conspiracy theory, and you have the Diebold CEO on the grassy knoll with a second USB stick and the paper ballot folds back and to the left, back and to the left...
posted by Cool Papa Bell at 10:44 AM on August 7, 2015 [1 favorite]


If the treatment assignment (which precints get voting machines) is correlated with anything that is not fraud that would affect the percentage change over time

And we know that some correlation exists because the intercepts for different machines are very different. In WI, for whatever reason strongly Democratic counties bought Optec machines while heavily Republican counties went for Dominion/Sequoia. I expect that's as simple as "Milwaukee County bought Optec and Waukesha bought Sequoia," though.
posted by ROU_Xenophobe at 10:53 AM on August 7, 2015


At some point, the cries of voter machine fraud start to resemble every other case of conspiracy theory

The difference is that it is ongoing. Year in and year out we vote black box systems when we don't have to, because the for-profit makers claim some for-profit reason why their systems shouldn't be verifiable and votes shouldn't be checkable nor meaningful recounts possible and apparently that's good enough for us.
posted by George_Spiggott at 11:01 AM on August 7, 2015 [1 favorite]


Also, there’s the fact that the polls are red-shifted (where there is systematic biasing toward Republican candidates) and have been for several election cycles. This is routinely assumed to be due to Republicans being less likely to answer pollsters, but there is no empirical justification for it. It’s used by polling organizations in their models in order to more accurately predict official results.
In the UK we call this the "Shy Tory" phenomenon, which I feels covers the situation more appropriately: right-wing voters ought to be ashamed of themselves; some are at least smart enough to act ashamed.


See, this was one of those signals to me that the author didn't know what they were talking about.

First, it would be the results, not the polls, that were red shifted according to their thesis. Second, the results have not been consistently red shifted — the best example of this was the "Unskewed Polls" nonsense of 2012. Third, the cause in 2014 of Republicans overperforming their poll results came not from Republicans being less likely to answer pollsters but rather the methods used to normalize poll results based on a model where Democrats are less likely to answer pollsters. Which is true — Dems are less likely to answer pollsters because they tend to be on cell phones (which are harder to get data for), tend to work more jobs off of the 9-5 schedule, and tend to screen their calls more. All of those biases are highly correlated to characteristics that are also highly correlated to partisan preference, e.g. age and income levels. So in actual terms, Democrats respond to polls less than they vote, so pollsters work with models that oversample Democrat demographics (i.e. including more black people than their proportional representation in the population) and also weight those samples (i.e. having a response from one young person count as 1.1 older person, as a made-up weight).

But simply put: If you don't know how actual polling works and the statistical assumptions underlying it, I'm not going to take your analysis of statistical deviations from that polling as credible.
posted by klangklangston at 11:20 AM on August 7, 2015 [2 favorites]


The difference is that it is ongoing.

So are claims of voter ID fraud, which we know are code for something else, and happening at such a low level as to be literally insignificant. So what's the point?

Year in and year out we vote black box systems

And yet where is the outcry for optical voting machines, which solve the problem completely? It seems the conversation is, "the bad people are doing bad things so let's all feel bad about this grarrrr."

As opposed to, "bad things can possibly happen, so here's an idea to do something else."

Otherwise, it's just Internet outrage, full of sound and fury, signifying nothing.

It's bad MetaFilter, really.
posted by Cool Papa Bell at 11:26 AM on August 7, 2015


And yet where is the outcry for optical voting machines, which solve the problem completely?

It's hard to raise an outcry unless there is something to cry out about?

I don't think most of the public even has the idea that those magical touch screens they use during elections might have programming underlying them, and the idea that the programming might not simply be an electronic abacus, even on their radar.
posted by hippybear at 11:29 AM on August 7, 2015 [1 favorite]


I don't think most of the public even has the idea that those magical touch screens they use during elections

Most voters use paper ballots that are optically scanned.
posted by ROU_Xenophobe at 11:32 AM on August 7, 2015


Otherwise, it's just Internet outrage, full of sound and fury, signifying nothing.

It's bad MetaFilter, really.


So don't talk about electronic voting machines on the internet, why?
posted by MisantropicPainforest at 11:40 AM on August 7, 2015


Yeah, what is the penetration of electronic voting machines across the US anyway?
posted by hippybear at 11:40 AM on August 7, 2015


I've worked as a poll clerk and just want to say, counting ballots manually is really not very difficult, takes only a few hours, and I'm not sure why expensive machines are necessary at all. I'm not sure what problem is being solved. I guess I don't really see them as any sort of improvement in terms of the potential to be manipulated.
posted by Hoopo at 11:48 AM on August 7, 2015 [2 favorites]


The irony here is that for the last umpteen years one party has been decreeing that voter fraud exists only as a method to attempt disenfranchise large groups of people. It is very difficult to mentally acknowledge that voter fraud probably does exist simply because I quite literally disagree with everything that particular party has stood for in the last 30 years or so.

So for me at least, it's damned if it's there and damned if it's not. I say we just get rid of the current system, including the gerrymandered lines all of the nonsensical over-legal laws. Get suggestions from every elementary school kid in the country, and build something from there. I don't care if we have to somehow come up with flying dragons and underwater school buses, those sound like better investments than anything involving what we have now.
posted by Blue_Villain at 11:52 AM on August 7, 2015


And yet where is the outcry for optical voting machines, which solve the problem completely?

Several States Abandon Electronic Voting for Paper
Maryland switches voting machines again
Why Are Some States Dumping Their Electronic Voting Machines and Going Back to Paper?

Plenty more where that came from. Most of these stories say something about "voter distrust of touch screen voting" and "despite millions spent on electronic voting". Doesn't seem like a lack of outcry to me.
posted by George_Spiggott at 11:52 AM on August 7, 2015 [3 favorites]


. It is very difficult to mentally acknowledge that voter fraud probably does exist simply because I quite literally disagree with everything that particular party has stood for in the last 30 years or so.

Voter fraud simply doesn't exist as a thing. We're talking a few fraudulent votes per billion, if that.

Electoral fraud, however...
posted by feckless fecal fear mongering at 11:57 AM on August 7, 2015 [4 favorites]


I have always voted on optical scan ballots, but were I ever to live in a touchscreen/no-paper-trail voting district I would without question request an absentee ballot for every election and make sure I was going to be out of the district on voting day if the law requires it. Software is hard to write under situations where there isn't a temptation to manipulate results. I'm just not going to trust a pure software implementation of the electoral process.
posted by ndfine at 12:34 PM on August 7, 2015


because of course making sure that your own individual vote is counted accurately is totes a solution to the problem of most people voting by insecure means.
posted by You Can't Tip a Buick at 12:40 PM on August 7, 2015


because of course making sure that your own individual vote is counted accurately is totes a solution to the problem of most people voting by insecure means.

I didn't make that claim and I don't think it's fair to read that into what I said.
posted by ndfine at 12:46 PM on August 7, 2015


Yeah, what is the penetration of electronic voting machines across the US anyway?

From memory: about 55% of votes are cast using optical scanners and about 40% using direct recording machines. I don't think I've ever seen a breakdown of how many direct-recorded votes were cast with VVPAT and without.
posted by ROU_Xenophobe at 12:46 PM on August 7, 2015


> I didn't make that claim and I don't think it's fair to read that into what I said.

Then what's the point of always personally getting an absentee ballot? I think I'm missing something here...
posted by You Can't Tip a Buick at 12:53 PM on August 7, 2015


Then what's the point of always personally getting an absentee ballot? I think I'm missing something here...

I agree that the system is broken and that while I can't fix everything, I can do something to help ensure my individual vote is accurately counted. That was my entirely hypothetical point and not that I was proposing a workable solution for the whole of our broken electoral system.

I would prefer a mail based ballot system ala OR and WA. I am ok with currently voting on optical scan ballots and would do what I can do stop a move to a purely electronic system were it to be pushed here. If it happens that I move to a jurisdiction with a purely electronic system, I would do what I can to encourage a move away from the electronic voting system, but in the short term, I would vote via absentee ballot.
posted by ndfine at 1:03 PM on August 7, 2015


It just seems like a waste of time and effort to me, especially when you get down to going through the trouble to arranging to be away on election day. I understand the attachment to the idea that a vote is an expression of individual will and that therefore you're protecting yourself by ensuring that your personal vote isn't changed, but given that elections are about collectively deciding on allocation of power rather than about expressing oneself individually, you wouldn't really be doing anything at all by making sure your personal vote isn't changed.

Maybe the hypothetical you in this scenario could organize as part of a group, all of whom arrange to be out of town on election day, and then try to draw media attention to that action? That might be worthwhile. But simply protecting your individual vote on an individual level is just sort of a fetishy game, rather than an actual political action, and may end up serving as a placebo for political action.

I mean, it fundamentally doesn't matter; we're like three layers deep in hypotheticals right now. The scenario you're proposing just struck me as something that looks like a protest, but isn't.
posted by You Can't Tip a Buick at 1:13 PM on August 7, 2015


It just seems like a waste of time and effort to me, especially when you get down to going through the trouble to arranging to be away on election day.

Would I literally leave the jurisdiction on election day? Probably not. If I recall, the laws for absentee ballot access differs across the states where it isn't required at all to being basically unenforceable where it is.

[S]imply protecting your individual vote on an individual level is just sort of a fetishy game, rather than an actual political action, and may end up serving as a placebo for political action.

It is less of a fetish when you're talking about things like small municipal elections. I am fully aware that elections are about collective action, but I have lived in jurisdictions where the whole electorate wasn't even 200 voters. I currently live in a municipality that had < 1500 voters turnout in the last election, and it's an inner ring suburb with a growing tax base in a metro area of 1+ million. The instinct to protect my individual vote, particularly in elections like those, does not come from a desire to protest nor as a placebo in lieu of political action. It's not large scale elections like senate or presidential elections where I particularly worry about manipulation of electronic voting systems--it's those tiny, off-year elections where a shift of a handful of voters can change results. Large amounts of money, policy, and power are controlled by elected officials in areas like that all over the country.
posted by ndfine at 2:35 PM on August 7, 2015


In precincts with electronic voting machines, are paper ballots also available for people who want them, or is one stuck with electronic voting?
posted by Spathe Cadet at 4:43 PM on August 7, 2015


As mentioned above, you can register as an absentee voter, in which case you get to cast a paper ballot.

The catch is, however, that the absentee votes are not counted AT ALL except when they might sway a vote.

And, well, a rigged electronic voting machine can certainly prevent that from happening.
posted by ocschwar at 6:09 PM on August 7, 2015 [1 favorite]


The catch is, however, that the absentee votes are not counted AT ALL except when they might sway a vote.

That is not true. All valid absentee ballots are counted in every election. Your absentee ballot may not be counted before an obvious loser concedes the election, but the official final tally always includes absentee ballots.

Many absentee ballots are rejected because they failed to be received on time, you used the wrong envelope or your signature doesn't match, but all valid absentee ballots are counted.
posted by JackFlash at 7:30 PM on August 7, 2015


I don't want this to be true. This truthiness makes me prejudiced against the evidence. I cant objectively evaluate it. So it's a conundrum. I'd like to see the data made available to a trusted third party like Nate Silver.
posted by humanfont at 9:29 PM on August 7, 2015


I agree with Tom Scott.
posted by flabdablet at 10:13 AM on August 8, 2015


I don't want this to be true. This truthiness makes me prejudiced against the evidence. I cant objectively evaluate it. So it's a conundrum. I'd like to see the data made available to a trusted third party like Nate Silver.

It's not that it's not true -- there's no reason to expect that she's wrong/lying about large precincts trending Republican. There's just no reason at all to connect that to electoral fraud. What she's picked up is virtually certain to be an artifact of counties choosing voting machines/methods (where there's variation in a state) and choosing how to divide themselves into precincts.
posted by ROU_Xenophobe at 12:32 PM on August 8, 2015


Alas, postal ballots are a problem for freedom to vote as we please.

All it takes is one overbearing individual at home to bully a household member into voting postally against their choice, and the whole system starts to look like an exercise in effective corruption. For this reason, postal ballots are as illegal as e-voting in the Netherlands.
posted by rum-soaked space hobo at 10:08 AM on August 9, 2015


All it takes is one overbearing individual at home to bully a household member into voting postally against their choice

My state's started issuing postal ballots to everyone, and this is definitely my main concern about that. No enforced ballot privacy means that it's relatively easy for someone with power over you (employer, family member, etc) to oversee your vote. Though we do have a method voters can use to void a previously-submitted vote and substitute a new one in person, I think that's probably more of a hassle than is reasonable for most people to do.

In spite of my misgivings, I think increasing the accessibility of voting offsets the lack of privacy. We don't even need to find postage stamps: there are special ballot drop boxes at every polling place. Being able to drop off ballots at any time of day or night is pretty great. I would like to see more public education about what expectations voters should have about the privacy of their ballots, as well as what recourse they have if that privacy has been compromised coercively (and if the laws in place for this aren't sufficient, then we should change them.)
posted by asperity at 12:30 PM on August 9, 2015


« Older "Women – Love each other, support each other...   |   Several secular Bangladeshi bloggers hacked to... Newer »


This thread has been archived and is closed to new comments