The curious case of Alvin Greene

[Galteeth]

So I'm sure you guys have heard about this.

I can believe that he won the election by fluke.

But what's his deal personally? The answers he gives in interviews are obviously coached, and are also obviously lies. Why? Ok, maybe some opportunist decided to become his manager and is giving him bad advice? But still, why lie?

Some people think he's a republican plant. But that doesn't add up either. Why would republicans waste their time in an election that's so safe for them? And if they were going to "plant" this guy, wouldn't they at least give him some kind of an advisor or something? Or encourage him to leave his parents' basement?

It's a fascinating story. Something seems off here, but what is it?

 

[ibnsos]

So I'm sure you guys have heard about this.



Some people think he's a republican plant. But that doesn't add up either. Why would republicans waste their time in an election that's so safe for them? And if they were going to "plant" this guy, wouldn't they at least give him some kind of an advisor or something? Or encourage him to leave his parents' basement?

It's a fascinating story. Something seems off here, but what is it?

What I don't understand is, even if he is a republican plant, the democrat voters still voted for him, so why would it matter who backed him. It's not like he even had a campaign.

 

[Ygggdrasil]

Statistical analysis from the vote counts in the South Carolina primary indicate some irregularities. Vic Rawl, the losing candidate in the primary, released a press release talking about the finding of two electroal forensic experts on the primary:

Dr. Mebane performed second-digit Benford's law tests on the precinct returns from the Senate race. The test compares the second digit of actual precinct vote totals to a known numeric distribution of data that results from election returns collected under normal conditions. If votes are added or subtracted from a candidate’s total, possibly due to error or fraud, Mebane’s test will detect a deviation from this distribution.

Results from Mebane’s test showed that Rawl’s Election Day vote totals depart from the expected distribution at 90% confidence. In other words, the observed vote pattern for Rawl could be expected to occur only about 10% of the time by chance. “The results may reflect corrupted vote counts, but they may also reflect the way turnout in the election covaried with the geographic distribution of the candidates' support,” Mebane said.

(http://www.vicrawl.com/vicrawl/post/1002-statement-by-the-vic-rawl-for-us-senate-campaign )

Tom Schaller, a statistician from fivethirtyeight.com adds:

Indeed: An unusual, non-random pattern in the precinct-level results suggests tampering, or at least machine malfunction, perhaps at the highest level. [...] Along with the anomalies between absentee ballot v. election day ballots Miller found, something smells here.

(http://www.fivethirtyeight.com/2010/06/sc-democratic-primary-getting-weirder.html)

Unfortunately, as SC's electronic voting systems leave no paper trail for auditing, it may be difficult to determine whether any fraud/malfunction played a role in the result.

 

[BobG]

Some people think he's a republican plant. But that doesn't add up either. Why would republicans waste their time in an election that's so safe for them? And if they were going to "plant" this guy, wouldn't they at least give him some kind of an advisor or something? Or encourage him to leave his parents' basement?

It's a fascinating story. Something seems off here, but what is it?

Lee Atwater would pull something like this just for the fun of it.

Fortunately most politicians have no shame or they might have been embarrassed to have their former campaign manager contract brain cancer, convert to Catholicism, and publically repent and apologize to each of the politicians he trashed during his past campaigns.

 

[Galteeth]

Statistical analysis from the vote counts in the South Carolina primary indicate some irregularities. Vic Rawl, the losing candidate in the primary, released a press release talking about the finding of two electroal forensic experts on the primary:
(http://www.vicrawl.com/vicrawl/post/1002-statement-by-the-vic-rawl-for-us-senate-campaign )

Tom Schaller, a statistician from fivethirtyeight.com adds:
(http://www.fivethirtyeight.com/2010/06/sc-democratic-primary-getting-weirder.html)

Unfortunately, as SC's electronic voting systems leave no paper trail for auditing, it may be difficult to determine whether any fraud/malfunction played a role in the result.

Hmm. Maybe some genius will show up and explain that he did this to show how easy it is to manipulate these e-voting machines. And then promptly be arrested.

In all seriousness, if we are talking about a margin of a ten percent swing, then perhaps fraud could be the explanation. But an unknown guy getting 60 percent through fraud? If that is really possible, and easy enough to pull off that it doesn't require a fairly large conspiracy, then our democracy is in real trouble.

It seems to me that the explanation of a confluence of 1) low turnout 2) Greene's name being in higher position on the ballot 3) Greene having a slightly more familiar sounding name 4) People voting against Rawls and 5) Republican cross overs voting for what they perceived as the unknown candidate, seems a more likely explanation.

Partially because again, what does the GOP really have to gain from committing such massive fraud in what is seen for them as one of the safest senate elections in the country? The risk/reward ratio doesn't make any sense. Furthermore, in South Carolina's GOP, there are bitter divisions, and someone would likely sing.

I know this is all speculation, but the only facts in this story just don't make sense.

As I said, the fluke win thing doesn't seem that unbelievable to me (as the statisticians pointed out, there is a one in ten chance of this happening randomly, which really isn't that unlikely, I would have thought it would be more) but something about Greene himself is such so dang odd. It seems like someone should investigate that lawyer person who was allegedly feeding him answers during the Olbermann interview (he can be heard off camera saying, "he's a candidate" when Greene gave the wikipedia-ish answer about his opponent).

On a slightly tangenital note, showing a college student porn is a felony? Really? Given the circumstances as described by the girl, that seems fairly draconian.

 

[Galteeth]

What I don't understand is, even if he is a republican plant, the democrat voters still voted for him, so why would it matter who backed him. It's not like he even had a campaign.

Strictly speaking, if there was no fraud, he's a democrat, and republicans gave him money/ voted for him, it probably wouldn't.

However, one thing I learned from working on the Ron Paul campaign is that primary election are not in the same field as national elections. In most cases, the rules governing the primary are set by the parties, and are not laws, and hence can be changed. There are some exceptions to this, as may be the case here, when federal election machinery is used, and the laws vary state to state, but more or less the parties aren't bound by hard rules when it comes to the nomination process. The actual election is a different matter.

 

[Galteeth]

Lee Atwater would pull something like this just for the fun of it.

Fortunately most politicians have no shame or they might have been embarrassed to have their former campaign manager contract brain cancer, convert to Catholicism, and publically repent and apologize to each of the politicians he trashed during his past campaigns.

If we want to get into real conspiracy territory, we could suggest that perhaps this was some sort of a "test run" by someone to see how much e-voting fraud they could get away with and not be detected.

Again, a far fetched explanation. I sort of find this story interesting because it's a good political mystery.

 

[BobG]

If we want to get into real conspiracy territory, we could suggest that perhaps this was some sort of a "test run" by someone to see how much e-voting fraud they could get away with and not be detected.

Again, a far fetched explanation. I sort of find this story interesting because it's a good political mystery.

I was referring to finding a way to encourage an embarrassing candidate to run in the other party's primary vs fraud.

And they probably did get lucky. Primaries are notorious for low voter turnout, meaning small organized groups can get their specialty candidate on the general election ballot, but it still isn't something that could be done clandestinely. The only problem with the getting lucky part is that people motivated enough to vote in primaries usually know who their candidates are.

Yeah, it's a mystery. Probably just luck, but some very unlikely luck.

 

[Vanadium 50]

Statistical analysis from the vote counts in the South Carolina primary indicate some irregularities.

Hardly. Nobody who approaches this as a scientist can possibly come to this conclusion.

#1: A 90% confidence level is very low indeed. At this threshold, in a world with perfect elections, we would declare about 47 Congressional elections fraudulent every cycle.

#2: The information comes from the losing candidate. Do you think that he is showing a random distribution of studies? Or do you think he is picking the one that makes his point best? If the latter, how many studies do you think he looked at before selecting the Mebane study? 2? 5? 10? 20? How significant is a 90% CL now?

#3: Mebane's study says "The value of j-hat for Rawl matches the value observed in Mebane (2010a; 2010b) for many losing legislative candidates in U.S. elections during the 1980s and 2000s, and so might not be considered all that unusual." That's not what the losing candidate's press release says Mebane's study is saying.

#4: Why would elections even follow Benford's Law at all? Benford's Law applies to systems with sizes that cover several orders of magnitude. Election precincts are set up to keep their sizes roughly the same.

 

[TheStatutoryApe]

#1: A 90% confidence level is very low indeed. At this threshold, in a world with perfect elections, we would declare about 47 Congressional elections fraudulent every cycle.

#2: The information comes from the losing candidate. Do you think that he is showing a random distribution of studies? Or do you think he is picking the one that makes his point best? If the latter, how many studies do you think he looked at before selecting the Mebane study? 2? 5? 10? 20? How significant is a 90% CL now?

I'm no math wiz so I would not be capable of unraveling the numbers but from the quote in the article it seems they are saying there is a 90% confidence level of tampering, or that there is only a 10% chance that the voting patterns would have turned out this way.

 

[Al68]

Anyone remember Fred Tuttle in Vermont? He ran for U.S. Senate in 1998 in the Republican primary, then democratic voters voted in the Republican primary to ensure Tuttle won the primary. Then as soon as he won the Republican primary, he went on TV to endorse and glorify his Democratic opponent, Patrick Leahy. They had a TV "debate" that consisted of Tuttle asking people to vote for his opponent, Leahy, and talking about how great Leahy was and how terrible he himself would be in office, while Leahy made his obligatory denouncement of democrats voting in the Republican primary.

If you think this sounds more like a comedy movie than reality, you're only half right. Before this fiasco, Tuttle was an actor who starred in a movie as a local yokel running for U.S. Representative for Vermont with a hilariously absurd campaign.

http://en.wikipedia.org/wiki/Fred_Tuttle

 

[Ygggdrasil]

#2: The information comes from the losing candidate. Do you think that he is showing a random distribution of studies? Or do you think he is picking the one that makes his point best? If the latter, how many studies do you think he looked at before selecting the Mebane study? 2? 5? 10? 20? How significant is a 90% CL now?

#3: Mebane's study says "The value of j-hat for Rawl matches the value observed in Mebane (2010a; 2010b) for many losing legislative candidates in U.S. elections during the 1980s and 2000s, and so might not be considered all that unusual." That's not what the losing candidate's press release says Mebane's study is saying.

These are valid criticisms.

#4: Why would elections even follow Benford's Law at all? Benford's Law applies to systems with sizes that cover several orders of magnitude. Election precincts are set up to keep their sizes roughly the same.

A conference paper looking at the 2nd digit Benford's Law test is available here. The paper does mention that its use as a method to detect election fraud is controversial, so this may be a valid criticism (I don't know enough about the test to say one way or the other).

#1: A 90% confidence level is very low indeed. At this threshold, in a world with perfect elections, we would declare about 47 Congressional elections fraudulent every cycle.

I'm no math wiz so I would not be capable of unraveling the numbers but from the quote in the article it seems they are saying there is a 90% confidence level of tampering, or that there is only a 10% chance that the voting patterns would have turned out this way.

Here's how you would interpret the result. According to Mebane, the 2nd digits from the vote counts should follow a certain distribution, derived from Benford's law. The results from the SC primary deviate significantly from the Benford's law distribution. Now, there are two possible explanations for this observed deviation:

1) The election results are valid and the deviation occurred by chance.
2) The election results are fraudulent.

(there may be other explanations, but for simplicity I'll just look at these two). Now, what is the probability that the SC results could occur by chance? If you assume that the 2nd digits of the vote counts are distributed according to Benford's law and generate millions of sets of vote counts from that theoretical distribution, only 10% of the results will show a deviation greater than the deviation shown by the 2010 primary results.

Now, to properly interpret these results, we should take a Bayesian approach. Let's say that one out of every thousand elections is fraudulent (this number is arbitrarily chosen for the purposes of this example, it is probably/hopefully an overestimation). This fraudulent election will fail the Benford's law test. However, of the other 999 elections, approximately 100 of them will also fail the Benford's Law. Therefore, of the 101 elections that fail the Benford law test, only 1 of them will have been fraudulent. Therefore, (given the assumed fraudulent election rate above) the likelihood that the election was fraudulent is still only about 1% (up from our prior probability of 0.1%). So, yes, given a rare event like a fraudulent election, even a test with a 90% confidence level does not give you much confidence that the election was fraudulent.

If the Benford's law test were the only piece of evidence we had, I would definitely agree with Vanadium. However, there are other pieces of evidence that point to possible fraud. One issue brought up by the fivethirtyeight post I mentioned previously, is that the voter turnout numbers for the Republican races look odd. There's also the large discrepancies between absentee vs election day ballots (although discrepancies in this case are to be expected). Finally, there's the surprise of a candidate who did not campaign at all winning the election.

I do agree that these data certainly are not conclusive evidence of fraud/voting machine malfunction, and it would be very hard to prove so. However, I don't think the possibility can be fully dismissed. While I think human manipulation of the vote count is unlikely, I do wonder what the likelihood of voting machine malfunction is. The manufacturer of the South Carolina voting machines has had problems before (http://techdirt.com/articles/20100609/1616099761.shtml). I guess a larger point that I'm trying to make here is that some of these issues could have been addressed if the voting machines included some sort of paper trail for auditing later on.

 

[Galteeth]

A somewhat related anecdote: I was the ron paul precinct captain for my precinct during the presidential primaries. Our precinct used diebold voting machines, and I was poll watching. At one point, they stopped using a machine because it had malfunctioned. They had a diebold employee come to the precinct, take the machine off location, and he returned a couple hours later saying the problem was fixed. Now, I don't know that any fraud occurred, and the voting results from that precinct matched the exit polls, but it was still an eye opening experience that nobody had a problem with a diebold employee removing a machine from the voting location and returning it.

 

[BobG]

Anyone remember Fred Tuttle in Vermont? He ran for U.S. Senate in 1998 in the Republican primary, then democratic voters voted in the Republican primary to ensure Tuttle won the primary. Then as soon as he won the Republican primary, he went on TV to endorse and glorify his Democratic opponent, Patrick Leahy. They had a TV "debate" that consisted of Tuttle asking people to vote for his opponent, Leahy, and talking about how great Leahy was and how terrible he himself would be in office, while Leahy made his obligatory denouncement of democrats voting in the Republican primary.

If you think this sounds more like a comedy movie than reality, you're only half right. Before this fiasco, Tuttle was an actor who starred in a movie as a local yokel running for U.S. Representative for Vermont with a hilariously absurd campaign.

http://en.wikipedia.org/wiki/Fred_Tuttle

Yeah, but Tuttle ran in the primary against a candidate that didn't even know how many teats a cow had! How in the world could that guy answer "six?" when asked that in the primary debate?!

And quite a few felt Tuttle's running was more to provide publicity for the movie he had appeared in than an actual attempt to win (or a conspiracy to disrupt the other party's election).

 

[Vanadium 50]

but from the quote in the article it seems they are saying there is a 90% confidence level of tampering, or that there is only a 10% chance that the voting patterns would have turned out this way.

That's what Rawl would like you to believe, but it's almost exactly the opposite of what a 90% confidence level means. A 90% CL means that you will miscategorize a fair election as dishonest 10% of the time. It makes no statement about how often one would miscategorize a dishonest election as fair.

 

[BobG]

One possible contributing factor is that South Carolina is a very conservative state. Four of the 6 Congressional districts are practically automatic Republican victories with one of the 2 Democratic districts being very conservative Democrats. They may have a Democratic Congressman, but that district almost always votes Republican for national elections.

There's only one truly Democratic district, created via gerrymandering and having the heaviest concentrations of black voters. Only three of the Congressional districts had contested Democratic Congressional primaries, with the 6th District (the Democratic one) having turnout 5 times greater than the other two, but the lack of numbers for comparison in the other 3 districts makes that an unreliable indicator. None the less, it shouldn't have come as a great surprise that a black would win a large number of votes in the 6th district.

States or districts represented by one party tend to be more likely to have some flaky candidates nominated for the general election, since sure defeat discourages many of your serious politicians.

My Congressional district is an automatic Republican victory. Not so bad that Democrats are putting up flakes for candidates, but they usually aren't very strong. The Libertarian Party, on the other hand, tends to nominate true flakes to run in the general election. They have this emphasis on body count, getting candidates in as many elections as possible, but that creates a self sustaining cycle where no one takes them serious. Last cycle of elections, two brothers that were unemployed elections each ran for a public office as a Libertarian candidate.

 

[jtbell]

Here's an update on the Alvin Greene story from the perspective of a South Carolina voter (me). He's been practically invisible during this campaign season, probably because of practically nonexistent resources. There's been very little about him on the TV news or in the newspapers since August, when he was indicted on felony obscenity charges for allegedly showing a USC student pornographic images in a campus computer lab. While driving around the Upstate, I haven't seen any campaign signs for him. Or for his opponent Jim DeMint, for that matter, who has probably the easiest re-election campaign of any US senator this year.

I did find this TV news report, from when he was stumping in person at the South Carolina State Fair last week:

https://www.youtube.com/watch?v=gDwWoD3GZrk