The only explanation, Mr. Rawl’s representatives told the committee, was faulty voting machines — not chance, name order on the ballot, or Republicans crossing over to vote for the weaker Democrat. With testimony dominated by talk of standard variances, preference theories and voting machine software, the hearing took on the spirit of a political science seminar.
What is your probability estimate for Alvin Greene’s win in this election being legitimate (Greene getting lucky as a result of aggregate voter intent+indifference+confusion, as opposed to voting machine malfunction or some sort of active conspiracy)? What evidence do you need in order to update your estimate?
Not ready to answer the rationalist questions, but why is it that, as soon as elections don’t go toward someone who played the standard political game, suddenly, “it must be a mistake somehow”? You guys set the terms of the primaries, you pick the voting machines. If you’re not ready to trust them before the election, the time to contest them was back then, not when you don’t like the result.
Where was Rawl on the important issue of voting machine reliability when they did “what they’re supposed to”?
I understand that elections are evidence, and given the prior on Greene, this particular election may be insufficient to justify a posterior that Greene has the most “support”, however defined. But elections also serve as a bright line to settle an issue. We could argue forever about who “really” has the most votes, but eventually we have to say who won, and elections are just as much about finality on that issue as they are as an evidential test of fact.
To an extent, then, it doesn’t matter that Greene didn’t “really” get the most votes. If you allow every election to be indefinitely contested until you’re convinced there’s no reason the loser really should have won, elections never settle anything. The price for indifference to voting procedure reliability (in this case, the machines) should be acceptance of a bad outcome for that time, to be corrected for the next election, or through the recall process.
Frankly, if Greene had lost but could present evidence of the strength Rawl presented, we wouldn’t even be having this conversation.
On election night, I was among the first reporters to speak with Greene after his victory was announced. His verbal tics and strange affect were immediately apparent: he frequently repeats and interrupts himself, speaks haltingly, and sometimes descends into incoherent rambling, as subsequent video and audio interviews have made all the more obvious.
Damn those candidates with autism symptoms! Only manipulative people like us deserve to win elections!
I should point out that most of the people who ought to know about the issue, have been screaming bloody murder about electronic voting machines for some time now. Politicians and the general public just haven’t been listening. This issue is surfacing now, not because it wasn’t an issue before, but because having a specific election to point to makes it easier to get people to listen. It also helps that the election wasn’t an important one (it was a Democratic primary for a safe Republican seat), and the candidates involved don’t have the resources to influence the discussion like they normally would.
This doesn’t sound like autism to me. It sounds more like a neurotypical individual who is dealing with a very unexpected and stressful set of events and having to talk about them.
Be that as it may, those are typical characteristics of high-functioning autistics, and I’m more than a little bothered that they view this as justification for reversing his victory.
Take the part I bolded and remove the “incoherent rambling” bit, and you could be describing me! Well, at least my normal mode of speech without deliberate self-adjustment.
And my lack of incoherent rambling is a judgment call ;-)
But elections also serve as a bright line to settle an issue. We could argue forever about who “really” has the most votes, but eventually we have to say who won, and elections are just as much about finality on that issue as they are as an evidential test of fact.
To an extent, then, it doesn’t matter that Greene didn’t “really” get the most votes. If you allow every election to be indefinitely contested until you’re convinced there’s no reason the loser really should have won, elections never settle anything.
Yes. Exactly. This is true for lawsuits as well: getting a final answer is more important than getting the “right” answer, which is why finality is an important judicial value that courts balance.
My most likely explanations would be 1) software bug(s) 2) voter whim or confusion 3) odd hypothesis no one has thought of yet. Active intent to steal the nomination a distant fourth. Make it 60⁄30 among the first two.
Evidence? Well, anything credible, but how likely is that. :)
I put a very high probability that some form of tampering occurred primarily due to the failure of the data to obey a generalized Benford’s law. Although a large amount of noise has been made about the the fact that some counties had more votes cast in the Republic governor’s race than reported turnout, I don’t see that as strong evidence of fraud since turnout levels in local elections are often based on the counting ability of the election volunteers who often aren’t very competent.
I’d give probability estimates very similar to those of Jim’s but with a slightly higher percentage for people actually voting for him. I’d do that I think by moving most of the probability mass from the idea of someone tampering with the election to expose the insecure voting machines which implies a very strange set of ethical thought processes. I’ve also had enough experience in local elections to know that sometimes very weird things happen for reasons that no one can explain (and that this occurs even with systems that are difficult to tamper with). So using the primary breakdown given by Jim I’d put it as follows:
* Voters actually voted for him: 0.25
* Someone tampered with the voting machines or memory cards to make Alvin Greene win: 0.25
* ...and that person did it because they wanted Alvin Greene to win: 0.1
* ...and that person did it for kicks: 0.1
* ...and that person did it because they wanted to expose the insecure voting machines: 0.05
* Someone meant to tamper with a different election on the same ballot, but accidentally altered the democratic primary additionally or instead: 0.1
* The votes were altered by leftover malware from a previous election which was also hacked: 0.2
* There was a legitimate error in setting up or managing the voting machines altered the vote totals: 0.2
Edit: Thinking this through another possibility that should be listed is deliberate Republican cross-over (since it is an open primary) but given the evidence that seems of negligible probability at this point (< .01)).
Edit: Thinking this through another possibility that should be listed is deliberate Republican cross-over (since it is an open primary) but given the evidence that seems of negligible probability at this point (< .01)).
I would count that under “voters actually voted for him”
I don’t know the details about the American voting system, but (or maybe therefore) I am surprised how low estimates all people give to the possibility that the result is genuine. My estimate (without much research, I’ve just read the links) is
0.5 voters actually voted for Greene
0.3 error of some kind
0.2 conspiracy
In order to update, any evidence is accepted, of course. What I would most like to see: results of some statistical survey, conducted either before or better after the election, historical data concerning performance of black candidates, historical data from elections with big difference between the intensity of the campaign between the competing candidates, a lot of independent testimonies of trustworthy voters reporting non-standard behaviour of the voting machines, description of how can the results be altered (and what is normally done to avoid that).
* 0.6 voters actually voted for Greene
* >0.3 error of some kind
* <0.1 conspiracy
Conspiracy is a really stupid claim for this result—it is an incredibly unimportant election. If someone was going to purposely jigger the results of an election, they would do it where it actually mattered. The only reason it is still on there is that people sometimes do do really stupid things (as opposed to normally stupid things that they do all the time).
Someone tampered with the voting machines or memory cards to make Alvin Greene win: 0.4
...and that person did it because they wanted Alvin Greene to win: 0.1
...and that person did it for kicks: 0.1
...and that person did it because they wanted to expose the insecure voting machines: 0.2
Someone meant to tamper with a different election on the same ballot, but accidentally altered the democratic primary additionally or instead: 0.1
The votes were altered by leftover malware from a previous election which was also hacked: 0.2
There was a legitimate error in setting up or managing the voting machines altered the vote totals: 0.2
Note that I started researching this topic with an atypically high prior probability for voting machine fraud, and believe that it is very likely that major US elections in the past were altered this way. The strongest direct evidence I see for fraud having occurred is that there were “three counties with more votes cast in Republican governor’s race than reported turnout in the Republican primary” FiveThirtyEight. Note that this means botched vote fraud, not correctly-implemented vote fraud, since correctly implemented vote fraud, using a strategy such as the Hursti hack, would have changed the votes but not the turnout numbers.
The Benford’s Law analysis on FiveThirtyEight, on the other hand, I find very unconvincing—first because it has a low p-value, and second because it doesn’t represent the way voting machine fraud really works; it can only detect if someone makes up vote totals from scratch, rather than adding to or subtracting from real vote totals.
That is not important when considering the probability that Alvin Greene would have a worse influence on the Senate than the average politician if he got elected. It is only important when considering the probability that he would have a much worse influence on the Senate than average.
I mean, in the sense that the US government is like a massive Ouija board that is not really controlled by anyone, then sure. But the senators seem to have a particularly heavy hand on the board.
Because, unless he is a politician, the sentence fails to make sense, because ‘more standard politician’ requires him to be one? If so, I think being selected as a candidate makes you a politician.
I think voters were clueless about both candidates, but they like to fill in all the boxes on the ballot, so they chose the name that has the higher positive affect by far: “Alvin Greene”.
To me that would be sufficient to explain the entire anomaly, if not for the mysterious origin of Greene’s $10,000 filing fee.
Why is this at all ridiculous? Is there any reason to believe Arnold Schwarzenegger has done a significantly worse job than other governors, controlling for ability of the legislature to agree on anything and the health of the economy?
It merely serves to illustrate what politics is really about. It certainly isn’t about voting for people who are the best suited for making and implementing the decisions that are best for the country, planet or species. I actually would have voted for him unless he had a particularly remarkable opponent. All else being equal I take a contribution in another field that is popular and that I appreciate is a more important signal to me than success as a pure courtier. It is unfortunate that I do not have reason to consider consider political popularity as a stronger signal of country-leading competence than creating ‘kindergarten cop’.
Is there any reason to believe Arnold Schwarzenegger has done a significantly poorer job than other governors, controlling for ability of the legislature to agree on anything and the health of the economy?
I’ve already assigned a low probability to Alvin being at all worse than the alternatives. I expect Arnold would be ‘even’ better.
(Oh, and I do think that one liner is sub par. It would be better to stick to actual ridiculous rather than superficially ridiculous.)
Strange occurrence in US South Carolina Democratic primary.
The Washington Post profiled Alvin Greene last week
10 minute video interview with Greene
What happened here?
Wikipedia has a list of possible explanations.
Fivethirtyeight lists possible explanations and analysis.
Rawl and co presented five hours of testimony that the results could only be attributed to a problem with the voting machines.
What is your probability estimate for Alvin Greene’s win in this election being legitimate (Greene getting lucky as a result of aggregate voter intent+indifference+confusion, as opposed to voting machine malfunction or some sort of active conspiracy)? What evidence do you need in order to update your estimate?
Not ready to answer the rationalist questions, but why is it that, as soon as elections don’t go toward someone who played the standard political game, suddenly, “it must be a mistake somehow”? You guys set the terms of the primaries, you pick the voting machines. If you’re not ready to trust them before the election, the time to contest them was back then, not when you don’t like the result.
Where was Rawl on the important issue of voting machine reliability when they did “what they’re supposed to”?
I understand that elections are evidence, and given the prior on Greene, this particular election may be insufficient to justify a posterior that Greene has the most “support”, however defined. But elections also serve as a bright line to settle an issue. We could argue forever about who “really” has the most votes, but eventually we have to say who won, and elections are just as much about finality on that issue as they are as an evidential test of fact.
To an extent, then, it doesn’t matter that Greene didn’t “really” get the most votes. If you allow every election to be indefinitely contested until you’re convinced there’s no reason the loser really should have won, elections never settle anything. The price for indifference to voting procedure reliability (in this case, the machines) should be acceptance of a bad outcome for that time, to be corrected for the next election, or through the recall process.
Frankly, if Greene had lost but could present evidence of the strength Rawl presented, we wouldn’t even be having this conversation.
ETA: Oh, and you gotta love this:
Damn those candidates with autism symptoms! Only manipulative people like us deserve to win elections!
I should point out that most of the people who ought to know about the issue, have been screaming bloody murder about electronic voting machines for some time now. Politicians and the general public just haven’t been listening. This issue is surfacing now, not because it wasn’t an issue before, but because having a specific election to point to makes it easier to get people to listen. It also helps that the election wasn’t an important one (it was a Democratic primary for a safe Republican seat), and the candidates involved don’t have the resources to influence the discussion like they normally would.
This doesn’t sound like autism to me. It sounds more like a neurotypical individual who is dealing with a very unexpected and stressful set of events and having to talk about them.
Be that as it may, those are typical characteristics of high-functioning autistics, and I’m more than a little bothered that they view this as justification for reversing his victory.
Take the part I bolded and remove the “incoherent rambling” bit, and you could be describing me! Well, at least my normal mode of speech without deliberate self-adjustment.
And my lack of incoherent rambling is a judgment call ;-)
Well… knowing that someone is autistic is some inferential evidence in favor of them being a good hacker.
Yes. Exactly. This is true for lawsuits as well: getting a final answer is more important than getting the “right” answer, which is why finality is an important judicial value that courts balance.
My most likely explanations would be 1) software bug(s) 2) voter whim or confusion 3) odd hypothesis no one has thought of yet. Active intent to steal the nomination a distant fourth. Make it 60⁄30 among the first two.
Evidence? Well, anything credible, but how likely is that. :)
I put a very high probability that some form of tampering occurred primarily due to the failure of the data to obey a generalized Benford’s law. Although a large amount of noise has been made about the the fact that some counties had more votes cast in the Republic governor’s race than reported turnout, I don’t see that as strong evidence of fraud since turnout levels in local elections are often based on the counting ability of the election volunteers who often aren’t very competent.
I’d give probability estimates very similar to those of Jim’s but with a slightly higher percentage for people actually voting for him. I’d do that I think by moving most of the probability mass from the idea of someone tampering with the election to expose the insecure voting machines which implies a very strange set of ethical thought processes. I’ve also had enough experience in local elections to know that sometimes very weird things happen for reasons that no one can explain (and that this occurs even with systems that are difficult to tamper with). So using the primary breakdown given by Jim I’d put it as follows:
Edit: Thinking this through another possibility that should be listed is deliberate Republican cross-over (since it is an open primary) but given the evidence that seems of negligible probability at this point (< .01)).
I would count that under “voters actually voted for him”
Ok. Yeah, so that should probably be a subcategory of that in that it explains the weird results in a sensible fashion.
I don’t know the details about the American voting system, but (or maybe therefore) I am surprised how low estimates all people give to the possibility that the result is genuine. My estimate (without much research, I’ve just read the links) is
0.5 voters actually voted for Greene
0.3 error of some kind
0.2 conspiracy
In order to update, any evidence is accepted, of course. What I would most like to see: results of some statistical survey, conducted either before or better after the election, historical data concerning performance of black candidates, historical data from elections with big difference between the intensity of the campaign between the competing candidates, a lot of independent testimonies of trustworthy voters reporting non-standard behaviour of the voting machines, description of how can the results be altered (and what is normally done to avoid that).
I would say
Conspiracy is a really stupid claim for this result—it is an incredibly unimportant election. If someone was going to purposely jigger the results of an election, they would do it where it actually mattered. The only reason it is still on there is that people sometimes do do really stupid things (as opposed to normally stupid things that they do all the time).
Here is my probability distribution:
Voters actually voted for him: 0.1
Someone tampered with the voting machines or memory cards to make Alvin Greene win: 0.4
...and that person did it because they wanted Alvin Greene to win: 0.1
...and that person did it for kicks: 0.1
...and that person did it because they wanted to expose the insecure voting machines: 0.2
Someone meant to tamper with a different election on the same ballot, but accidentally altered the democratic primary additionally or instead: 0.1
The votes were altered by leftover malware from a previous election which was also hacked: 0.2
There was a legitimate error in setting up or managing the voting machines altered the vote totals: 0.2
Note that I started researching this topic with an atypically high prior probability for voting machine fraud, and believe that it is very likely that major US elections in the past were altered this way. The strongest direct evidence I see for fraud having occurred is that there were “three counties with more votes cast in Republican governor’s race than reported turnout in the Republican primary” FiveThirtyEight. Note that this means botched vote fraud, not correctly-implemented vote fraud, since correctly implemented vote fraud, using a strategy such as the Hursti hack, would have changed the votes but not the turnout numbers.
The Benford’s Law analysis on FiveThirtyEight, on the other hand, I find very unconvincing—first because it has a low p-value, and second because it doesn’t represent the way voting machine fraud really works; it can only detect if someone makes up vote totals from scratch, rather than adding to or subtracting from real vote totals.
Probability that this person would have a worse influence on the senate than a more standard politician: 5%.
I would give it lower than that, US Senators have surprisingly little power.
That is not important when considering the probability that Alvin Greene would have a worse influence on the Senate than the average politician if he got elected. It is only important when considering the probability that he would have a much worse influence on the Senate than average.
???
I mean, in the sense that the US government is like a massive Ouija board that is not really controlled by anyone, then sure. But the senators seem to have a particularly heavy hand on the board.
Sorry, I meant “influence”, not “power”.
Conditional on their winning the election, presumably.
I’m not sure that is technically necessary given the precise phrasing.
Because, unless he is a politician, the sentence fails to make sense, because ‘more standard politician’ requires him to be one? If so, I think being selected as a candidate makes you a politician.
It seems to make sense without any fancy interpretation.
I think voters were clueless about both candidates, but they like to fill in all the boxes on the ballot, so they chose the name that has the higher positive affect by far: “Alvin Greene”.
To me that would be sufficient to explain the entire anomaly, if not for the mysterious origin of Greene’s $10,000 filing fee.
Also the possible “Al Green” effect—voters may have thought they were voting for the famous soul singer.
The next election being won by a ficus would boost my estimate. Or, you know, something else ridiculous like an action hero actor.
Why is this at all ridiculous? Is there any reason to believe Arnold Schwarzenegger has done a significantly worse job than other governors, controlling for ability of the legislature to agree on anything and the health of the economy?
It merely serves to illustrate what politics is really about. It certainly isn’t about voting for people who are the best suited for making and implementing the decisions that are best for the country, planet or species. I actually would have voted for him unless he had a particularly remarkable opponent. All else being equal I take a contribution in another field that is popular and that I appreciate is a more important signal to me than success as a pure courtier. It is unfortunate that I do not have reason to consider consider political popularity as a stronger signal of country-leading competence than creating ‘kindergarten cop’.
I’ve already assigned a low probability to Alvin being at all worse than the alternatives. I expect Arnold would be ‘even’ better.
(Oh, and I do think that one liner is sub par. It would be better to stick to actual ridiculous rather than superficially ridiculous.)
Probability that this person would have a worse influence on the senate than a more standard politician: 5%.
my breakdown:
Conspiracy: 19%
Error of some kind: 80%
Voters actually voted for him: 1%
(Given that there is a unique cause, and it is one of those three, of course.)
Surely the “voters aren’t actually paying attention” hypothesis deserves more than 1% probability.
that could fall under ‘error of some kind’.