We were on the road all day long yesterday, after getting the morning’s Green News Report. We finally got off the road and back on the grid as the sun was setting here in the Midwest, and as “results” began coming in from some of the five different states (Alaska, Arizona, Florida, Oklahoma and Vermont) holding primary elections last night.
It seemed a good moment to ask a question on Twitter about what we’d missed so far. In the bargain, we received some interesting Twitter replies from NBC’s Chief White House Correspondent and election results guy Chuck Todd, in regard to counting ballots and the fact that he (and others) were busy reporting “results” to the world, even as not one single ballot had yet to be counted by any human being in all of the United States by that point.
Here’s how some of that conversation on Twitter went…
(For those not familiar with Twitter, “RT” means “ReTweet”. It’s akin to quoting the person you are responding to.)
@TheBradBlog:
(As Todd had been tweeting various incoming results, we decided to ask him a direct question about them.)
@TheBradBlog:
(And here, just a few minutes later, after receiving SDWinkler’s specious, unsupported “myth” tweet, alleging that human counts are not more accurate than machine counts, Todd appears to do a 100% flip-flop on his previous “Great Q” position.)
@ChuckTodd:
(…Before he then asks us the following question.)
@ChuckTodd:
(Todd doesn’t answer the above. But, if only momentarily, he becomes a bit more careful in his tweets declaring who is reportedly up or down, based on machine counts, in Florida’s tight Republican Gubernatorial primary between Rick Scott and Bill McCollumn.)
@ChuckTodd:
(Not long after polls closed in Arizona, former Washington Post, now Slate blogger Dave Weigel, who’d been similarly tweeting unverified, machine-counted “results” previously, gets curious as to why Arizona’s numbers — a state where they’ve had serious problems with voting machines in the past — are not coming out fast enough.)
@DaveWeigel:
(We then go back to tweeting to Todd.)
@TheBradBlog:
(Todd doesn’t respond to the above, and his “cautiousness” doesn’t last long, as a few minutes later, he tweets the following.)
@ChuckTodd:
(To which we reply…)
@TheBradBlog:
(A very few minutes later, Weigel, who also tweeted AP calling it for Scott, gets irritated at Bill McCollum for not conceding after the Republican Governors Association issues a statement congratulating Scott. That, despite the fact that ZERO votes have actually been counted by any human being in the entire state of Florida, and just a few thousand, unverified, machine-counted votes are said to separate McCollum from Scott.)
@DaveWeigel:
(Talking Points Memo’s Eric Kleefeld makes a wry joke, as he seems to get similarly antsy while waiting for Arizona to report on their unverified, machine-counted “results”.)
@EricKleefeld:
(Hoping to inform Kleefeld, Weigel and Todd about just some of the past problems with machine-counted votes in Arizona, we tweet the following.)
@TheBradBlog:
(No replies come to the above, before NBC’s Todd retweets a report of Arizona’s GOP Senate primary “results” to his 57,217 Twitter followers.)
@ChuckTodd:
At this hour, as we prepare to hit the road again for, largely, the entirety of the day, there remain several high-profile races where the media have not yet been able to declare a “winner”. Among them, the five-way race for the Democratic nomination for Governor in Vermont where, as reported by Politico late last night,
“140 Vermont towns count their ballots by hand, while 106 towns use optical scanners.”
The op-scanners used in most of Vermont were made by Diebold, and are serviced by a less-than-reputable (to say the least) company called LHS Associates. Diebold’s op-scanners were infamously hacked in HBO’s Emmy-nominated documentary Hacking Democracy. You can watch that landmark hack, which succeeds in flipping the results of an election undetectably (unless someone actually bothers to count the paper ballots by hand), “live” as it happened right here.
According to AP a few minutes ago, one candidate in the Vermont race has declared himself the winner, even though the reported results have him just 200 votes ahead of the second place candidate, and a few more than that in front of the third place candidate (who happens to be the Secretary of State).
In Alaska, the incumbent Republican U.S. Senator Lisa Murkowski is said to be “on the ropes” in a surprisingly close primary race against lesser known Joe Miller (who was endorsed by Sarah Palin). Miller is reportedly winning by 1,960 votes out of some 89,000 votes cast as we write this. Those votes were likely “counted” by the state’s Diebold optical-scan system. Alaska has featured a number of notoriously questionable elections results over the last few cycles.
The results from some of the rural areas in Alaska where, Anchorage Daily News reports this morning, “paper ballots are counted by hand” are not yet included in the reported results. But the final results, they say, “won’t be known for over a week. The Alaska Division of Elections said over 16,000 absentee ballots were requested and as of Monday night 7,600 had been returned. The first count of absentees will be next Tuesday and there will be two subsequent counts as the absentee votes trickle in on Sept. 3 and on Sept. 8.” Those absentee ballots will most likely be counted by Diebold’s hackable paper-based optical-scan systems as well
Whatever Diebold reports those results to be, reporters like Todd, Weigel and Kleefeld will most likely report them uncritically, no matter that it’s likely not even one of the ballots will have been examined by hand to assure that Diebold’s op-scan systems are “counting” them correctly. Hopefully Murkowski will demand that somebody does so before she concedes the Republican primary race. As always, we’re more than happy to offer her advice on how to do that, if she wishes to contact us.
brad,
you should of asked todd if franken is the senator from minn or not because if i remember correctly,the machine count gave coleman the victory,while the hand count gave us senator franken
Chuck Todd is a know-nothing douchebag. I think this should be every media outlet’s headline for a day. I think his media credentials came from a crackerjack box.
Brad,
Great post, though I take some issue with some of the implied message. Your “conversation” with Chuck Todd is absolutely typical of any interaction with mainstream reporters. It raises the question (a good one) of whether Chuck is unable to follow your clear logic or simply unwilling to engage in the obvious consequences of it (my vote is for the latter). Reporters in general are not paid to really think and are unwilling to question the rules of “the game”. Most of them actually prefer thinking of elections as a game. Even those who really do think about it prefer to yell about outcomes rather than worry about the details of how the numbers were obtained. That’s far more commercial.
Where I take some issue is with the supposition that hand counts are always preferable. Machines, if properly protected and supervised, are clearly faster and more accurate. As you point out, the problem arises because our voting machines are not properly protected and supervised. The right way forward is not to go back to a nation of 300 million voting with paper and pencil (which is also not tamper proof), but to require adequate protection and supervision. BTW, supervision means a voter-verifiable paper record that serves as the actual legal ballot and must be secured and retained for use in recounts and for press scrutiny.
good point, questioneverything.
todd is a paid shill…why did he go in two seconds from nothing to WH correspondent…he is supported by insiders…anyone that uses that ATM is purposely misinforming people, he is not that stupid, tho he often acts it
If it weren’t for our BF, I’d also say that “tweeter = douchebag.” How can a serious exchange take place with that nonsense? I’d no sooner have a twitter account than I would have an unanaethstetized vasectomy. However the “twit” part of it seems appropriate. I say this partly because I am a public school teacher whose students can’t hold together a cohesive argument about anything of substance.
@mark #3: Hand counting of hand-marked paper ballots may not be as perfectly “accurate” as machine counting, but it would be MUCH harder to steal an election.
And, the evidence would be there.
What’s the cost of an ATM vs Voting Machines and tabulators? Cost to make one, that is, I’m sur some precincts have overpaid for those POS’s. Actually, I guess every precinct has now that I think about it. I’m assuming a lot more goes into an ATM.
Mark, I don’t think machine counting is faster than hand counting. I think hand counting would be faster, especially considering the results are much more trustworthy.
Mark,
If you support retaining paper ballots for recounts, isn’t that a tacit admission that, when all is said and done, if you REALLY want to know who won an election, you count paper ballots by hand.
Yes? No? If not, what am I missing here? Why even keep paper ballots if you think machine counting is better?
???
I would add to my previous post that a “voter verifiable paper record” is not the same as a paper ballot and is also vulnerable to manipulation and fraud. Nevertheless, if you really believe that “machine is better,” than what’s up with keeping any paper at all?
I say, if you’re gonna keep paper, keep an original, human-filled-out paper ballot and COUNT the damn thing. By hand!
The fallacy in the statement “Human counting is not more accurate than machine counting” begins with the unprovable assumption that machines actually count.
In the case of DREs it is virtually impossible to verify whether the machines have so much as engaged in counting — as opposed to simply spewing out a predetermined result.
In the case of optical scans, it is virtually impossible to verify whether the system has counted “any” ballots, let alone counted them accurately without resorting to a hand count of all the ballots.
Therefore, in asserting “Human counting is not more accurate than machine counting” sdwinkler unloaded a load of crap and Todd either swallowed it without a second thought or he found, in sdwinkler’s load of crap the perfect Orwellian substance for feeding Brad Friedman’s direct question down a “memory hole” — (the device used in George Orwell’s 1984 to incinerate inconvenient facts).
Thanks for staying on message, Brad.
I’m with the “hand count is the only way” crowd. I don’t see that speed of results matters, no matter if you can get ballots counted faster by machine. I think we should all agree that accuracy is paramount. I also think that there are a number of places in the world where hand counts get done in 4 to 5 hours after the polls close. And with unlimited observation – anybody can watch the count. That’s what I wish we had here in TN. We have to trust our computer voting machines with no paper at all.
Mark @ 3 said:
And your evidence for that is what exactly?
In New Hampshire, where 40% of the precincts count by hand, many of the hand-counted areas often have their numbers in before the machine-counted areas.
Where do you get the idea that they’re not as fast? And if they are not as fast, one can simply add more counters, no?
As to accuracy, I’m unaware of any studies at all that indicate hand-counting is less accurate than machine counting. As far as I’ve seen, hand-counting is far MORE accurate than machine counting which is why, as I’ve noted in the past (and Lora notes above), when you really want to know who really won an election, you count the ballots by hand. Hand-counting is Democracy’s Gold Standard, so why settle for less?
Wrong. As I’ve pointed out many times on this blog, but must keep doing so apparently, the problem is not whether the votes are adequately protected. It’s not even whether they are counted accurately or not. It’s whether the citizens can know that they’ve been counted accurately.
If they can’t oversee the process, so that they know the election was recorded accurately, as per the intent of the voters, it’s as much of a threat to democracy as manipulated elections themselves.
Unless hand-marked paper ballots are actually counted by human beings, I’m unaware of any way that I can know that an election was counted accurately.
I still remain open to alternatives, I just know of none. And believe me, I’ve been looking. For YEARS.
BTW, supervision means a voter-verifiable paper record that serves as the actual legal ballot and must be secured and retained for use in recounts and for press scrutiny.
There is no bigger tool out there in the so-called non-Faux MSM today than Chuck Todd. I literally change the station anytime he comes on.
Which Steve is this?
Brad I know this is probably a very gauche question, but what I’m wondering is:
Why can’t we have a voting machine that counts electronically, and then spits out a reciept for the voter and for the counters?
That way there are paper ballots to be counted in the event of a recount request, and the voter also gets to hold their lil voter reciept and KNOW their vote counted the way they wanted it to.
I’m the (?original) “Steve” who has been posting here since 2004. Definitely not the troll “Steve” who has been posting here recently. It is confusing and I wish there was a way to differentiate (other than by what one posts). I don’t really want to change the name I use here (and in real life).
The problem I see with that is cleverly written software can still manipulate what is presented to you. Yes, you clicked candidate X like you wanted, and yes your receipt shows you chose candidate X, but the tabulator software in the machine or the central tabulator says you voted for candidate Z. Unless there is a physical hand count of the results to compare to the tabulator count you guess is a good as mine on who actually received the most legitimate votes. We have no way to be sure… Its all a guessing game and we are at the mercy of corporations who put their profit margins first.
@HankyDub
We cannot assume that the electronic “counting” and what got printed out will match. Computers can be programmed to show a different result through each communication channel.
The answer is not to bifurcate how results are reported. There must only be one paper record that gets reviewed–the paper ballot marked by the voter.
– Tom
from todays news,
Aug. 25) — A foreign intelligence agency breached the U.S. military’s classified computer network using a virus spread by an infected thumb drive, a top Pentagon official has revealed.
“It began when an infected flash drive was inserted into a U.S. military laptop at a base in the Middle East,” Deputy Defense Secretary William Lynn writes in an article in the latest issue of Foreign Affairs. “The flash drive’s malicious computer code, placed there by a foreign intelligence agency, uploaded itself onto a network run by the U.S. Central Command.”
that was for the “properly protected” crowd
if the pentagon can’t protect its computers,what makes you people think the voting computers can be protected?
and as for the..why not a machine that prints a paper receipt….what are you gonna do,ask every1 to bring back their receipt? or are you gonna put the receipt in a box to count them….if so why bother with the 6-7 grande machine?
another thing to ask chuck tood would be..what about the counts that go backwards as brad proved in monroe county?,,,,if his car odometor began going backwards would he think he hit a time warp or what?
Brad Friedman wrote:
Never mind accuracy. Are you aware of any study which establishes that a DRE has “counted” any votes in any election?
Spewing out predetermined percentages and numbers is not “counting.” If there is no way to “prove” that a result was or was not the product of a malicious hack, and no access to source codes, there is no scientific means to establish that what the machine did was engage in “counting,” period!
Steve @18 wrote:
Perhaps the confusion could be eliminated for everyone else if you two changed your handles to “original Steve” and “troll Steve.”
Ernie wrote:
No, but…we need to be careful here, since we’re conflating DREs (Direct Recording Electronic, usually touch-screen) voting machines and paper-based optical scan systems.
For clarity then:
* DREs are 100% unverifiable, with or without a so-called “Voter Verifiable Paper Audit Trail” (VVPAT). The VVPAT is a joke, useless, and provides no actual verifiability for anything. DREs can never be used in any system which wishes to call itself a democracy. Period.
* Paper-based optical-scan systems use paper-ballots, theoretically, to determine reported results. Those results come after the paper-ballot (either verifiably hand-marked by a voter or unverifiably printed by a computer) has been scanned through the machine. Trouble is, there is no way to know that the op-scan system has read the paper ballot correctly and/or that the op-scan system hasn’t been tampered with to report inaccurate results, unless one counts the paper ballots by hand — thus rendering the op-scan system useless in the first place.
So when folks talk about speed/accuracy of hand-counting versus machine-counting, at least in this thread, I’ll give the general benefit of the doubt that they’re talking about op-scan, rather than DRE. If they’re talking about DRE, we don’t even need to have a discussion, because that is not democracy at all, period. If they’re talking about op-scan, they can try to argue that it’s faster or more accurate, though I’m unaware of any actual legitimate information to back up that claim, even as ballots “counted” by op-scan may or may not be “counted” accurately at all. Only a hand-count of those ballots tell us whether they were or weren’t, and so we’re back to the obvious square one. “Obvious”, of course, to those of us who wish to recognize it, anyway.
brad said,
while one can not prove if the op scans are tampered with or just inaccurate…one can prove they are inaccurrate,for example in florida today on brevard county’s election site the machines have reported for precinct 216 palm bay firestation #4…7 registered voters and 16 cards cast for a 228% turn out
if you wonder how rick scott can get the primary win,just look at how he can turn out the votes…. in same 216 precinct the site reports 6 registered voters and scott gets 8 votes
if you look further(still in precinct 216)the dem race for senate shows no registered votes but 7 votes and the repub state district 29 shows 7 registered voters and 14 votes …so thats actually 21 cards cast or a 300% turn out
impossible = inaccurate
http://www.votebrevard.com/
Human counting of ballots is definitely more accurate than machine counting!
Why? It is well known that people indicate their intention to vote for a candidate by making a check mark or an X or even circling thee name of the candidate. Most optical scanners cannot recognize any of those so records “no vote” instead of the clear intention of the voter.
If you want a fast approximation of the vote, scan the ballots. But, when the candidates differ by a small percentage, the hand count is the ONLY way to be sure the INTENT of the voter is counted.
The media like fast counts. That’s why they don’t like hand counts. But, optical scan won’t read every vote correctly.
I don’t want a fast approximation of the vote. I want a hand-counted, accurate total to be the very first total I ever hear.
If hand counting is taking too long, then there aren’t enough people doing it.
Voting results need to be audited.
Auditing consists of hand-checking a small percent of randomly selected ballots against the reported results. If the hand-check does not report the same percentage results (within a small margin) as the reported results, a recount can be required.
Audits increase voter confidence in elections, discourage fraud (because it is likely to be detected), and detect voting machine error.
Let’s hear it for audits in every election!
Eleanor Hare said:
No. Let’s hear it for counting every damned ballot that every damned citizen has given their time to cast. Period.
“Audits” are yet another red-herring meant to keep us from doing what we ought to be doing: Casting votes and COUNTING them. Period.
Brad, I’m in South Carolina. We use the totally unreliable iVotronic DREs. And, you know how the Elections Commission does a “recount” of the few paper ballots (absentee, etc.) they have? They run them through the scanner again!
This is the “Chuck Todd” myth in action.
Hankydub #17 wrote,
I’m sure many folks here have a better grasp of this issue than I do, but FWIW, here’s my take.
First off, what would your receipt have on it? Your vote, of course, and what else? Not your name, I hope, because there goes the secret ballot.
Some kind of randomly assigned number, perhaps, or a bar code? Still, somewhere there would be a link to your name, right? So someone could probably figure out who you were and how you voted.
OK, no link. Just a receipt with your vote and a random number that’s truly untraceable back to your name (if such a thing exists).
Then what? A list of all the votes with their numbers next to them? You clutch your receipt in hand and go check your vote. I suppose if you can show your receipt (and it’s proven to be a true receipt — how, I’m not sure, and there could be counterfeiters perhaps printing up fake receipts, couldn’t there?) and show that the vote was incorrectly recorded, theoretically it could be corrected.
OK, but what about everybody else who voted? What percent of voters will check their votes? What about votes that were incorrectly recorded and nobody checks?
And with a receipt that you get to keep, whoever wanted to buy/coerce your vote can now demand to see your little receipt to make sure you voted the way they intended.
It is not enough to have hand counts, those counts MUST have standards that define a legal vote, and those standards MUST be followed. Otherwise, hand counting is no better that a machine counting.
“But evidence in the record here suggests that a different order of disparity obtains under rules for determining a voter’s intent that have been applied (and could continue to be applied) to identical types of ballots used in identical brands of machines and exhibiting identical physical characteristics (such as “hanging” or “dimpled” chads). See, e.g., Tr., at 238″”242 (Dec. 2″”3, 2000) (testimony of Palm Beach County Canvassing Board Chairman Judge Charles Burton describing varying standards applied to imperfectly punched ballots in Palm Beach County during precertification manual recount); id., at 497″”500 (similarly describing varying standards applied in Miami-Dade County); Tr. of Hearing 8 10 (Dec. 8, 2000) (soliciting from county canvassing boards proposed protocols for determining voters’ intent but declining to provide a precise, uniform standard). I can conceive of no legitimate state interest served by these differing treatments of the expressions of voters’ fundamental rights. The differences appear wholly arbitrary.”
http://www.law.cornell.edu/supc...0-949.ZD1.html
o/t, problem, test
To original Steve: Steve if you type something like ‘n/a’ or ‘x’ in the “Website” line of the comment sign in your name in the/your comment header turns red instead of being black. I think newcomers should alter their names if they are the same as established, frequent posters on a website but that’s just me. Using the “Website” line would allow you to retain Steve: red Steve = original Steve.
sholud be in above post: ‘sign-in, your’. See, my name has been changing color.
counts have been to ballot by hand (my opinion, sure)