Thoughts about Watergate, corruption, electoral fraud and skepticism.

[Kevin_Lowe]

This is my thread for abstract discussion of relatively philosophical issues about political shenanigans. I would prefer people keep trolls, flames and ad hominems to other threads. I can only Ignore offenders, but I'd be pleased if people managed to firewall this thread from the unpleasant ones.

Anyway.

I was thumbing through my copy of All The President's Men, since discussion in another thread had turned to Watergate as a point of comparison to the ongoing election issues, when a particular bit struck me.

The five men arrested at 2:30am had been dressed in business suits and all had worn Playtex rubber surgical gloves. Police had seized a walkie-talkie, 40 rolls of unexposed film, two 35-millimeter cameras, lock picks, pen-size tear-gas guns, and bugging devices that apparently were capable of picking up both telephone and room conversations.

Then a little later:

The tallest of the suspects, who had given his name as James W. McCord, Jr., was asked to step forward. He was balding, with a large, flat nose, a square jaw, perfect teeth and a benign expression that seemed incongruous with his hard-edged features.

The Judge asked his occupation.

"Security consultant", he replied

The Judge asked where.

McCord, in a soft drawl, said that he had recently retired from government service. Woodward moved to the front row and leaned forward.

"Where in government?" asked the Judge.

"CIA", McCord whispered.

The Judge flinched slightly.

Okay. Here comes the philosophical question.

This stuff actually happened.

However suppose some blogger wrote: "My theory about the Ohio/Wherever election is that probably a bunch of ex-CIA operators in business suits, wearing surgical gloves and equipped with a variety of specialised tools for breaking and entering and for compromising computers, broke into the place where they keep the tabulating computers at 2:30am and did a number on them".

I'm not putting myself forward for the JREF challenge when I say that I predict that several of the regulars here would laugh themselves sick at the sheer implausibility of that theory.

So what should count as a kooky theory and what should not, given that the actual events of Watergate sound so much like a Hollywood potboiler to modern skeptics? Is it actually skeptical to take it for granted that "secured" electronic voting gear was safe because it was under lock and key, or that because it would take a skilled operator on the inside to crock electoral software that it therefore is not a possibility that demands investigation?

At one extreme, is there any sensible place to stop short of wearing an alfoil hat, once we start thinking that teams of ex-CIA infiltration experts with James Bond gear might be sneaking around up to no good? On the other hand, are we at the mercy of such people and their friends if we don't make some allowance for the ugly fact that such people exist and might get up to no good?

That's enough from me. Over to you. What's the skeptical approach to such scenarios?

 

[a_unique_person]

This would be just about as kooky, but nowhere near as well known, for some reason.

 

[toddjh]

However suppose some blogger wrote: "My theory about the Ohio/Wherever election is that probably a bunch of ex-CIA operators in business suits, wearing surgical gloves and equipped with a variety of specialised tools for breaking and entering and for compromising computers, broke into the place where they keep the tabulating computers at 2:30am and did a number on them".

...

So what should count as a kooky theory and what should not, given that the actual events of Watergate sound so much like a Hollywood potboiler to modern skeptics?

The difference, as in all matters skeptical, is evidence. The evidence that the Watergate break-in occurred is compelling. The evidence that "ex-CIA operators in business suits, wearing surgical gloves and equipped with a variety of specialised tools for breaking and entering and for compromising computers, broke into the place where they keep the tabulating computers at 2:30am and did a number on them" is nonexistent.

Is it actually skeptical to take it for granted that "secured" electronic voting gear was safe because it was under lock and key, or that because it would take a skilled operator on the inside to crock electoral software that it therefore is not a possibility that demands investigation?

My impression is that you're talking about Ohio specifically, correct? But why limit yourself to one state? Isn't it possible that mysterious ex-CIA operators in business suits tampered with the votes in every state? Why is one a "possibility that demands investigation," but not the other? We need to deal in probabilities, not possibilities.

On the other hand, are we at the mercy of such people and their friends if we don't make some allowance for the ugly fact that such people exist and might get up to no good?

That's enough from me. Over to you. What's the skeptical approach to such scenarios?

An investigation is worthwhile if the probable benefits outweigh the known costs. Is there sufficient evidence of vote-tampering beforehand to believe that an investigation would be worth the cost in terms of tax dollars and further muddying of the political process?

Naturally, I don't have any numbers to give you, but I believe that is the question that needs to be asked.

Jeremy

 

[geni]

Re: Re: Thoughts about Watergate, corruption, electoral fraud and skepticism.

My impression is that you're talking about Ohio specifically, correct? But why limit yourself to one state? Isn't it possible that mysterious ex-CIA operators in business suits tampered with the votes in every state? Why is one a "possibility that demands investigation," but not the other? We need to deal in probabilities, not possibilities.

Because no one is going to be stubid enough to bother rigging the votes in a state where you are garenteed to win.

 

[toddjh]

Re: Re: Re: Thoughts about Watergate, corruption, electoral fraud and skepticism.

Because no one is going to be stubid enough to bother rigging the votes in a state where you are garenteed to win.

Who says only one party would be trying it?

Jeremy

 

[The Central Scrutinizer]

However suppose some blogger wrote: "My theory about the Ohio/Wherever election is that probably a bunch of ex-CIA operators in business suits, wearing surgical gloves and equipped with a variety of specialised tools for breaking and entering and for compromising computers, broke into the place where they keep the tabulating computers at 2:30am and did a number on them".

Skeptics would ask for evidence that a crime was even commited, instead of pointing to an affidavit that someone's brother's uncle's first cousin's best friend's ex-wife's nephew saw someone breaking into an office.

And you will notice that nowhere in your example do the words "and the media isn't reporting it" appear. That is the first sign of a loony conspiracy theory. See the "Ohio voting fraud" as an example.

You will also notice that nowhere in your example do the words "it is expensive to maintain a blog to report this story that the media is covering up, so please send donations" appear. See the "Ohio voting fraud" as an example.

 

[Silicon]

The problem is, anyone can conjure up any number of such scenarios for every facet of life. I could conjure up at least a thousand scenarios that sound no more implausible than the Watergate break in.

Where does it stop, if you have no evidence?

I say, you shouldn't begin unless you have evidence.

 

[Kevin_Lowe]

The difference, as in all matters skeptical, is evidence. The evidence that the Watergate break-in occurred is compelling. The evidence that "ex-CIA operators in business suits, wearing surgical gloves and equipped with a variety of specialised tools for breaking and entering and for compromising computers, broke into the place where they keep the tabulating computers at 2:30am and did a number on them" is nonexistent.

We agree completely so far, for most definitions of "evidence".

Maybe I should try to put the question another way to clarify what I was getting at.

Suppose the 2008 presidential election is between A and B, and according to all pre-election opinion polls and all exit polls the people of Foo County are divided 50/50, and voter turnout is historically about 50%. However when the paperless voting machines put their heads together, they conclude that the people of Foo County broke 55/45 for A, and 95% of them turned out to vote.

Non-partisan career academics examine the available evidence and conclude, although they differ on details, that the odds of this discrepancy being due purely to chance are less than 1%. However, polls do go wrong sometimes, albeit very rarely. A security analysis reveals that all the important bits of kit were continuously observed by multiple observers from both parties at all times, except between midnight on election day and 6am the day after, when the kit was under lock and key.

So the two available hypotheses are that the pre- and post-election polls were erroneous, and voters turned out in unprecedented numbers, or that someone got at the central server software before the election and also got at the election kit while it was under lock and key, which would take skilled and organised people with special equipment.

Obviously the available evidence doesn't allow you to conclude anything either way. I think the single most relevant question is whether you think this situation calls for further investigation, or whether the reasonable course is to assume it was just a fluke and do nothing unless further evidence of shenanigans falls into our hands of its own accord.

Followup questions for those who say "do nothing" would be whether an investigation is warranted for electoral results that are unlikely to a 0.1% level, or 0.01% and so on, or for prima facie impossible results (like turnouts over 100%).

My impression is that you're talking about Ohio specifically, correct?

Absolutely incorrect. I am deliberately trying to avoid too strong a parallel to any one place or incident.

But why limit yourself to one state? Isn't it possible that mysterious ex-CIA operators in business suits tampered with the votes in every state? Why is one a "possibility that demands investigation," but not the other? We need to deal in probabilities, not possibilities.

These are all valid questions. Since there were a number of suspicious break-ins before Watergate (I forget if anyone ever put their hand up to them) it does seem to be a bit of an "opening the floodgates" hypothesis. If Mission Impossible shenanigans took place once it's nigh certain they took place several times at least. I think this may be why people want to file them under "lunatic conspiracy theories", since it's hard to justify entertaining the possibility that such events are one-offs. You more or less commit yourself logically to a position which can be caricatured as full-tilt lunacy once you even open the door by a crack.

An investigation is worthwhile if the probable benefits outweigh the known costs. Is there sufficient evidence of vote-tampering beforehand to believe that an investigation would be worth the cost in terms of tax dollars and further muddying of the political process?

I agree, although I personally think that the idea that investigations "muddy the political process" is a thoroughly pernicious sound bite. Investigations, properly conducted, clarify affairs and lay to rest suspicions. Or confirm them. Either way, muddiness is not the problem.

Naturally, I don't have any numbers to give you, but I believe that is the question that needs to be asked.

It's certainly a very important question. The answer does hang entirely on whether or not you consider an investigation that clears up all suspicions as a benefit or a total waste of money. As you can probably guess I lean towards the first opinion, since I'm in the "better safe than sorry" camp when it comes to political corruption.

I think some people are running the line that we should hesitate to conduct investigations because if they don't find wrongdoing they will have been a "waste" - that's a position I find baffling. It might be the case that these people feel investigations are highly unlikely to find anything, but I'm not sure how they are making that judgement.

 

[Daylight]

But what if both sides are suppressing the evidence? One side so they don’t get caught doing the evil deed, the other side because they want to use evidence of the evil deed in the future against them.

 

[a_unique_person]

But what if both sides are suppressing the evidence? One side so they don’t get caught doing the evil deed, the other side because they want to use evidence of the evil deed in the future against them.

More like they already have the dirt on them for something else. Tammany Hall had a long and proud history, for example.

 

[a_unique_person]

Re: Re: Thoughts about Watergate, corruption, electoral fraud and skepticism.

The difference, as in all matters skeptical, is evidence. The evidence that the Watergate break-in occurred is compelling. The evidence that "ex-CIA operators in business suits, wearing surgical gloves and equipped with a variety of specialised tools for breaking and entering and for compromising computers, broke into the place where they keep the tabulating computers at 2:30am and did a number on them" is nonexistent.



My impression is that you're talking about Ohio specifically, correct? But why limit yourself to one state? Isn't it possible that mysterious ex-CIA operators in business suits tampered with the votes in every state? Why is one a "possibility that demands investigation," but not the other? We need to deal in probabilities, not possibilities.



An investigation is worthwhile if the probable benefits outweigh the known costs. Is there sufficient evidence of vote-tampering beforehand to believe that an investigation would be worth the cost in terms of tax dollars and further muddying of the political process?

Naturally, I don't have any numbers to give you, but I believe that is the question that needs to be asked.

Jeremy

Would you accept Nugan Hand as fact?

 

[toddjh]

Re: Re: Re: Thoughts about Watergate, corruption, electoral fraud and skepticism.

Would you accept Nugan Hand as fact?

I'm afraid I don't know enough to say, one way or the other.

Jeremy

 

[toddjh]

Suppose the 2008 presidential election is between A and B, and according to all pre-election opinion polls and all exit polls the people of Foo County are divided 50/50, and voter turnout is historically about 50%. However when the paperless voting machines put their heads together, they conclude that the people of Foo County broke 55/45 for A, and 95% of them turned out to vote.

Exit polls consistently favored Kerry in many states, only to have Bush sneak ahead in the final count. What statistics (even those backed up by "non-partisan career academics") can't address are the psychological factors that lead people who voted for one candidate to talk to exit pollsters more than people who voted for the other. I don't believe this is something that can be predicted or controlled for, and renders exit poll discrepancies -- especially those under 10% -- pretty weak as far as evidence goes.

Obviously the available evidence doesn't allow you to conclude anything either way. I think the single most relevant question is whether you think this situation calls for further investigation, or whether the reasonable course is to assume it was just a fluke and do nothing unless further evidence of shenanigans falls into our hands of its own accord.

I don't believe it was a "fluke." Call it a cautionary tale about the evils of relying on non-scientific polls that can't address self-selection bias.

Followup questions for those who say "do nothing" would be whether an investigation is warranted for electoral results that are unlikely to a 0.1% level, or 0.01% and so on, or for prima facie impossible results (like turnouts over 100%).

I don't believe it's possible to compute a probability of tampering from exit poll results alone. It's not as simple as determining confidence intervals, because human psychology is affecting the results.

If the results seemed to have more votes than there were registered voters in that precinct, then yes, I would consider that reason enough to investigate.

You more or less commit yourself logically to a position which can be caricatured as full-tilt lunacy once you even open the door by a crack.

Which is why it's important to have solid evidence that some wrongdoing took place at all before you put on the tinfoil hat.

I agree, although I personally think that the idea that investigations "muddy the political process" is a thoroughly pernicious sound bite. Investigations, properly conducted, clarify affairs and lay to rest suspicions. Or confirm them. Either way, muddiness is not the problem.

Mere mention of the word "Florida" was enough to elicit groans from both sides before last year's election. The endless recounts, lawsuits, and proclamations from election officials in 2000 did nothing to clarify anything, but made an awful lot of people lose confidence in the system (and not in a constructive way).

It might be the case that these people feel investigations are highly unlikely to find anything, but I'm not sure how they are making that judgement.

Barring solid evidence, there's no reason to suspect any tampering, and thus no reason to investigate. Unless you're proposing investigating everything, in which case we're back to a colossal waste of resources.

Jeremy

 

[Random]

This entire discussion is why I am firmly against any sort of paperless voting system. If there are paper ballots, you at least have the option of physically counting them. Almost two-thirds of Florida’s 2004 votes exist purely as ones and zeros in voting machines manufactured by companies that support the Republican Party. I have no evidence that there was any computer vote fraud in Florida last election, but I can never be sure. Ever. We can’t even look at the insides of the voting machines because they may contain proprietary information.

Of course, having one side or the other steal an election through electronic vote fraud is not the nightmare scenario as far as our democracy is concerned (as horrible as that would be). The nightmare scenario is we start getting the computer tallies in on election night and the results are wrong. Fundamentally wrong. A million votes for Bush in a town with a hundred voters, twenty percent of the votes going to a candidate named “hakerz rule!”, and the electoral college being won by Buchanan (James, not Pat).

(edited for clarity)

 

[Kevin_Lowe]

Exit polls consistently favored Kerry in many states, only to have Bush sneak ahead in the final count. What statistics (even those backed up by "non-partisan career academics") can't address are the psychological factors that lead people who voted for one candidate to talk to exit pollsters more than people who voted for the other. I don't believe this is something that can be predicted or controlled for, and renders exit poll discrepancies -- especially those under 10% -- pretty weak as far as evidence goes.

For the sake of the example I'll ask you to believe that the exit pollsters in question learned from the mistakes of the past and that statistical estimates of probability like "this error has a less than 1% probability of being due to chance" are mathematically solid.

I don't believe it was a "fluke." Call it a cautionary tale about the evils of relying on non-scientific polls that can't address self-selection bias.

I don't believe it's possible to compute a probability of tampering from exit poll results alone. It's not as simple as determining confidence intervals, because human psychology is affecting the results.

I'll talk about this particular issue at more length at the very end of this post, but I strongly prefer not to let discussion degenerate into a just a discussion of the merits of exit polls.

If the results seemed to have more votes than there were registered voters in that precinct, then yes, I would consider that reason enough to investigate.

Okay. In that case, is there any way a good skeptic can catch electoral fraud in proprietary, paperless voting systems unless the perpetrators goof? Or do we just have to accept that fruad will take place and we will not catch it unless the criminals in question do something silly like run up a prima facie impossible vote count?

Which is why it's important to have solid evidence that some wrongdoing took place at all before you put on the tinfoil hat.

Okay. Here's where I think the problem with this line of thinking becomes clear.

Suppose there are very few instances of leopards eating people in Fooland. In fact while there have been suspicious disappearances, and there are as many leopards now in Fooland as there have ever been, there has not been a confirmed incident of leopards eating people since 1972.

A child disappears in the jungle. There is no hard evidence it was eaten by a leopard, but neither is there anything to disprove the leopard theory. (From time to time children do disappear in Fooland and the cases remain unsolved).

However in Fooland, the idea that someone might have been eaten by a leopard is a "tinfoil hat" idea. Anyone who suggests the possibility that a leopard was responsible for anything is ridiculed unless they have indisputable evidence it really was a leopard. Especially if they compound their ridiculousness by suggesting that if this most recent child was a leopard victim, it would logically be reasonably likely that some previously unexplained disappearances were also leopard attacks.

As a result, in Fooland no one ever thoroughly checks into child disappearances. After all, you wouldn't want people to think you believed it might have been a leopard. That would be ridiculous. Leopards haven't eaten anyone since 1972.

Mere mention of the word "Florida" was enough to elicit groans from both sides before last year's election. The endless recounts, lawsuits, and proclamations from election officials in 2000 did nothing to clarify anything, but made an awful lot of people lose confidence in the system (and not in a constructive way).

I don't think that the public attention span should be the limiting factor in investigations of corruption and fraud. I also think you are pointing the finger in the wrong direction.

The results of the system were indeed a bit of a mess. Gore would have won Florida on a full recount, Gore won the popular vote, Gore would have won Florida without vote suppression and misleading ballots, Bush would have won Florida on the partial recount that was all Gore could afford, Bush won the election under the rules as they stood, and Bush would have won the election under the rules as they stood even without Florida since Republicans controlled parliament.

The problem isn't that the results were contested and examined, though. The muddiness was generated by partisans on both sides making misleading statements about the results in a muddy torrent, then and ever since, and by the fact that the Florida election system genuinely was a corrupt mess. You are blaming the guy who lifted the floorboards and found the termite infestation for the damage done by the termites, and that's backwards.

I'm also not sure how you get to the conclusion that people lost confidence in the system and not in a contructive way, unless your test for being constructive is ensuring that something is done about the problems in which case I agree. Losing false confidence is a system is always a good thing.

Barring solid evidence, there's no reason to suspect any tampering, and thus no reason to investigate. Unless you're proposing investigating everything, in which case we're back to a colossal waste of resources.

I don't think those statements follow from each other.

Barring solid evidence of tampering there may be less than solid evidence of tampering. If I find crumbs on the bench that is not solid evidence that my girlfriend made a sandwich. However it most certainly is reason to suspect the possibility that my girlfriend made a sandwich.

Now the difference between the sandwich case and the electoral fraud case is that a lot more hangs on the fraud case, and the consequences of not clearing up the fraud case are much more serious than the consequences of not clearing up the sandwich question.

There is a difference between investigating everything, and investigating everything that gives one grounds for suspicion that the fundamental mechanisms of democracy might have been tampered with. I'm actually not against "wasting" large sums of public money to make absolutely sure no doubt exists about the legitimacy of the process.

If you USians waste a few tens of millions of dollars every four hundred years or so, when Presidential elections show a one-in-a-hundred random discrepancy, I think that would be money well spent. Your mileage may vary.

ADDENDUM ON EXIT POLLS
It does depend on the poll - the German exit polls that get constantly referred to, and the Utah student-conducted exit polls, have always been accurate to within 1% from what I can recall reading. They use proportionally larger samples than Mitofsky used, but not by a huge amount, and they had better response rates, but again not by a huge amount.

The claim that discrepancies under 10% are "weak" evidence of something amiss is not statistically supportable except for a very broad definition of "weak", when it comes to a poll with 13000 respondents (the approximate total number of people Mitofsky surveyed nationwide). Such polls, in theory and mostly in practise, should be accurate within 1%.

I'm not picking on you in particular, but I strongly suspect that backwards reasoning is involved in a lot of the recent criticism of exit polls. People decide first of all that there is no evidence of fraud, then they cast about for reasons why they can dismiss the exit polls to support this. Before the election there was none of this dismissal of exit polls, and immediately after the election at least one Republican commentator said in print that the discrepancies were too big to be chance and were thus evidence of pro-Kerry bias in the pollster's employers. Everyone agreed that exit polls were the bees knees, until they started to be used as evidence that there had been a leopard attack. Then they suddenly weren't all that accurate after all.

Let's get real, people. Leopards haven't eaten anyone since 1972.

 

[toddjh]

For the sake of the example I'll ask you to believe that the exit pollsters in question learned from the mistakes of the past and that statistical estimates of probability like "this error has a less than 1% probability of being due to chance" are mathematically solid.

Your hypothetical now no longer has anything at all to do with the 2004 U.S. election, but yes, given that kind of information, I think an investigation would be justified.

Okay. In that case, is there any way a good skeptic can catch electoral fraud in proprietary, paperless voting systems unless the perpetrators goof?

No, not really. That's why any voting system must include a paper trail.

(From time to time children do disappear in Fooland and the cases remain unsolved).

Sounds like the people of Fooland should keep better tabs on their children.

Jeremy

 

[The Central Scrutinizer]

Re: Re: Re: Thoughts about Watergate, corruption, electoral fraud and skepticism.

Would you accept Nugan Hand as fact?

Vote Nugan Hand for President in 2008!!!!

 

[a_unique_person]

Re: Re: Re: Re: Thoughts about Watergate, corruption, electoral fraud and skepticism.

Vote Nugan Hand for President in 2008!!!!

They were two people. Did you read the article?

 

[The Central Scrutinizer]

Re: Re: Re: Re: Re: Thoughts about Watergate, corruption, electoral fraud and skeptic

They were two people. Did you read the article?

No.

How about Nugan for President, and Hand for VP?

 

[Kevin_Lowe]

This entire discussion is why I am firmly against any sort of paperless voting system. If there are paper ballots, you at least have the option of physically counting them. Almost two-thirds of Florida’s 2004 votes exist purely as ones and zeros in voting machines manufactured by companies that support the Republican Party. I have no evidence that there was any computer vote fraud in Florida last election, but I can never be sure. Ever. We can’t even look at the insides of the voting machines because they may contain proprietary information.

Precisely. Moreover, if we are as conservative about what constitutes grounds for an investigation as most USians seem to be (on the 'net at least) we can't ever have sufficient grounds to look at the insides of the machines either. The election is hermetically sealed by technological and philosophical barriers to investigation.

Of course, having one side or the other steal an election through electronic vote fraud is not the nightmare scenario as far as our democracy is concerned (as horrible as that would be). The nightmare scenario is we start getting the computer tallies in on election night and the results are wrong. Fundamentally wrong. A million votes for Bush in a town with a hundred voters, twenty percent of the votes going to a candidate named “hakerz rule!”, and the electoral college being won by Buchanan (James, not Pat).

That's not a nightmare scenario so much as a hilarious scenario.

A nightmare scenario is one party (and it does not matter one whit which one) having an eternal lock on all branches of government and the election machines, winning every election by 3-5%, forever, with no exit polls and no possibility of an audit.