And then there is the issue of fraud... even if your voting system is perfect, you need to account for fraud. At least one dead guy voted in that election:)
The US Supreme Court actually stopped the statewide "machine rejected" recount with the 7-2 decision. It was the immediate use of the certified number that ran along partisan lines.
In any case, I realize I probably should not have brought up Florida, because my proposal would be a law, and so would no be immune from a meddling court.
Actually, there was a third possibility. The Florida Supreme Court had already intervened twice, kind of making up the rules as they went along. First, they arbitrarily extended a recount deadline that had been Florida law (that is what the US Supreme Court decided 7-2 was wrong). Then they decided to recount all of the ballots statewide that had been rejected by the machines, despite there being no provision for this in Florida Law... in their defense, this decision was only stopped by a more partisan 5-4 vote.
So my thesis is that, unchecked, the Florida Supreme Court could very well have decided the election.
It was a tie. The first count was low enough to automatically trigger a recount under Florida law. So far, so good. The recount left a 300-vote margin out of millions cast. Then the absentees were added and we got closer to 1000. Then more votes were found and we got back down under 700. OK. Then we got to hand-recounting only in selected counties, again per Florida law. Still fine. Then the Florida Supreme Court jumped in and extended a statutory deadline. Hmm... Then counties started missing even the extended deadline, reverting to the original number. Then Gore challenged this original (now certified) number. Then the Florida State Supreme court ordered hand ballot counting to start again for ballots that were rejected by the machines. At this point, the US Supreme Court injected itself, and you know the result.
The State Supreme Court quickly cast aside the state's own laws - almost immediately.
Can you imagine the chaos if it all hung on a statistical model?
I'm not asking for a "statistical model". I'm asking for a codified definition of a tie. The use of statistics is limited to agreeing upon how precise the voting system is. The numbers can be decided up-front. They already had numbers like this codified to trigger a recount. I'm arguing that if a second count does not produce a larger margin then you need to call it a tie and not pretend that the system is more precise than it is. There would have been no "chaos" because the election would have been a tie, and the backup method would have been implemented. In Washington State it apparently is a coin toss. It could just as well be tossed to the legislature.
But the key points are that the tools in the law will be made to work, and adding a statistical procedure to dictate that a tie has occured really does nothing to improve the situation,
I think that the definition of an exact numerical tie needs to be more broad, that's all. Recounts won't ever make a 2,000,001 to 199,999,999 election valid.
Just the one... I don't have a 5-digit account number or anything, but mine's pretty low:)
I agree that the electoral college is consistent with the way the constitution was written. I was just arguing that it isn't only a means to deal with poor communication - it also served to coax low-population states into union with large-population states. The constitution is full of such compromises.
I would like a simple "select whoever is acceptable to you" ballot. As many checkboxes as you would like - put one next to everyone you can stand the sight of. I believe it is called Approval Voting, and I think it is simple enough for regular folks to grasp.
It's not just a relic of antique communication systems... it also represented a compromise which gave low-population states more influence than a straight-popular vote would have afforded them.
That is not a federal government problem to solve.
I agree. The shame of the 2000 election was that Florida was so paralyzed by a tie that the US Supreme Court had to step in. This is exactly what I would like to prevent in the future.
After all, when things are THAT close, you really can't be seriously wrong by choosing either,
That is probably the best point made on here. Essentially, why should I care who wins after a tie? I actually agree, except that revolutions have been started for less. The pretense of legitimacy must be upheld.
I think an election result that is too close to call should be called a tie. Then the process can move on. I think it is silly if the law defines a tie as 2,000,000 to 2,000,000 but not 2,000,001 to 199,999,999. That's the mentality that leads to the Florida election in 2000. In that case, the initial margin was something like 400 votes. Out of millions. At that point, the law should have declared the margin too narrow to declare a victor and the next step in the process should kick in - be it a recount, a decision by the legislature, or even a coin flip.
I'm not asking for hocus pocus. I'm asking for a simple measure that honestly describes how accurate the process is. If you don't meet that threshold, you go to Plan B. Recounts are a joke if the numbers you are talking about are lower than the precision of the voting/counting method. Just decide the election with a coin flip and stop the charade if that is your plan.
A statistical analysis would NOT have helped. Nothing you've suggested would have helped.
I can't see how you look at Florida and declare success. The US Supreme Court had to fly in and save the day (or ruin the day). The whole process was a colossal fuck-up.
A state with millions of votes was trying to come up with an "answer" when a few hundred ballots separated the winner and loser. You will never get an answer, because the answer is that it was a tie... as you said, the voters had no clear preference.
So go with Plan B, and that can be whatever you want it to be. Another election, refer the matter to the legislature, whatever. Just don't let it get to the point where the state abdicates it's duty to decide who won. There was no process that allowed for a tie, so the courts had to step in. That should be unacceptable, but we don't like to admit that an election can be a tie and we don't define the criteria for a tie.
Yes, we do have a great method to decide how to handle a tie - the problem is that we don't have a mechanism to declare an individual state election to be a tie.
If you are getting hung up on the semantics of me calling it margin of error, then call it something else. Precision works for me. The important thing is to accept that the system is not perfect and that every recount will also have errors (or will also be imprecise). At some point, all that will happen is random results as the number of votes delta between the candidates is below the precision of the voting method. You could recount or you could flip a coin - the result would be the same. At some point you have to declare the election a tie and move to a backup mechanism.
I wasn't arguing against a paper ballot or a simplified ballot - they both seem like good ideas to me.
I'm arguing that the election should be declared a "tie" if the margin falls within (for instance) the 2% error. Then backup solutions could kick in. The frustrating part is that there exist backup systems, but today they never get triggered because no one codified what a spoiled election looks like.
In 2000 in the US, when the Florida machines counted a 400-vote margin, it was immediately clear that the election was a statistical tie. But instead of an orderly process, there were these pointless recounts until the Supreme Court finally decided to step in. Luckily, the US is stable enough that the government didn't have a major crisis. A president more like Nixon and less like Clinton could have easily made things much more exciting.
That's kind of my point - that there can in fact be an election that is a "tie"... a second type of result in addition to the current "winner" result. At that point, backup mechanisms could kick in so you don't get stuck with the Supreme Court jumping in to what should have been an orderly handoff to the legislature.
I was screaming back in 2000 that the whole quagmire in Florida could have been avoided if the Constitution had simply been followed.
The courts got involved because there was disagreement about whether or not the election was valid. In my opinion the election was quite obviously a statistical tie, and you are correct that the existing backup mechanism could have sorted it out. The problem is that there was no predefined criteria for an election that was simply too close to call.
No. I'm talking about having a clearly defined system for resolving an ambiguous result. In other words, not just hoping that the government holds together when the Supreme Court decides to step in and impose such a system after the fact.
So, yes - one possibility is to hold the election again - perhaps as they do in much of the world where they narrow the field to the two top candidates. Another is to allow the legislature to decide. Another is a coin flip. It doesn't matter - just so the process is clearly defined and considered legitimate.
And the election is not a statistical analysis of the votes of the population, it is the actual votes.
If you are telling me with a straight face that millions of ballots are counted with no mistakes, I have a bridge to sell you.
For an election to work the system must be held to a standard of infallibility and that all errors fall on the voter,
That's impossible. No system is infallible.
if it's found not to be the case and is significant to have possibly affected the outcome a re-election is called.
Which is exactly what I am calling for. A set of criteria, that if met, declares the result to be invalid. I'm just asking for the rules to be clearly defined in the beginning.
So the margin of error is factored in
No it isn't. You said so yourself - the system is presumed to be infallible.
There is a fundamental flaw in elections today: lack of consideration for "margin of error". In my opinion, margin of error should be calculated and any election which falls within the margin of error should either be held again or some sort of tie breaker should kick in.
Pretending that we can deduce the intention of every voter with zero errors is noble, naive, and ridiculous.
Slashdot Mirror
Or for the polling clerk to tally the wrong column. Or to zone out for a moment and forget to record one.
But even more likely is that a few dead people will vote :)
That's a pretty thorough system.
But still an error can occur.
And then there is the issue of fraud... even if your voting system is perfect, you need to account for fraud. At least one dead guy voted in that election :)
Sorry, I got that a bit wrong.
The US Supreme Court actually stopped the statewide "machine rejected" recount with the 7-2 decision. It was the immediate use of the certified number that ran along partisan lines.
In any case, I realize I probably should not have brought up Florida, because my proposal would be a law, and so would no be immune from a meddling court.
Actually, there was a third possibility. The Florida Supreme Court had already intervened twice, kind of making up the rules as they went along. First, they arbitrarily extended a recount deadline that had been Florida law (that is what the US Supreme Court decided 7-2 was wrong). Then they decided to recount all of the ballots statewide that had been rejected by the machines, despite there being no provision for this in Florida Law... in their defense, this decision was only stopped by a more partisan 5-4 vote.
So my thesis is that, unchecked, the Florida Supreme Court could very well have decided the election.
It actually WASN'T a tie, and you know that.
It was a tie. The first count was low enough to automatically trigger a recount under Florida law. So far, so good. The recount left a 300-vote margin out of millions cast. Then the absentees were added and we got closer to 1000. Then more votes were found and we got back down under 700. OK. Then we got to hand-recounting only in selected counties, again per Florida law. Still fine. Then the Florida Supreme Court jumped in and extended a statutory deadline. Hmm... Then counties started missing even the extended deadline, reverting to the original number. Then Gore challenged this original (now certified) number. Then the Florida State Supreme court ordered hand ballot counting to start again for ballots that were rejected by the machines. At this point, the US Supreme Court injected itself, and you know the result.
The State Supreme Court quickly cast aside the state's own laws - almost immediately.
Can you imagine the chaos if it all hung on a statistical model?
I'm not asking for a "statistical model". I'm asking for a codified definition of a tie. The use of statistics is limited to agreeing upon how precise the voting system is. The numbers can be decided up-front. They already had numbers like this codified to trigger a recount. I'm arguing that if a second count does not produce a larger margin then you need to call it a tie and not pretend that the system is more precise than it is. There would have been no "chaos" because the election would have been a tie, and the backup method would have been implemented. In Washington State it apparently is a coin toss. It could just as well be tossed to the legislature.
But the key points are that the tools in the law will be made to work, and adding a statistical procedure to dictate that a tie has occured really does nothing to improve the situation,
I think that the definition of an exact numerical tie needs to be more broad, that's all. Recounts won't ever make a 2,000,001 to 199,999,999 election valid.
Just the one... I don't have a 5-digit account number or anything, but mine's pretty low :)
I agree that the electoral college is consistent with the way the constitution was written. I was just arguing that it isn't only a means to deal with poor communication - it also served to coax low-population states into union with large-population states. The constitution is full of such compromises.
I would like a simple "select whoever is acceptable to you" ballot. As many checkboxes as you would like - put one next to everyone you can stand the sight of. I believe it is called Approval Voting, and I think it is simple enough for regular folks to grasp.
It's not just a relic of antique communication systems... it also represented a compromise which gave low-population states more influence than a straight-popular vote would have afforded them.
That is not a federal government problem to solve.
I agree. The shame of the 2000 election was that Florida was so paralyzed by a tie that the US Supreme Court had to step in. This is exactly what I would like to prevent in the future.
You probably know about it, but I find this interesting as well.
I also support improving our education system :)
After all, when things are THAT close, you really can't be seriously wrong by choosing either,
That is probably the best point made on here. Essentially, why should I care who wins after a tie? I actually agree, except that revolutions have been started for less. The pretense of legitimacy must be upheld.
I think an election result that is too close to call should be called a tie. Then the process can move on. I think it is silly if the law defines a tie as 2,000,000 to 2,000,000 but not 2,000,001 to 199,999,999. That's the mentality that leads to the Florida election in 2000. In that case, the initial margin was something like 400 votes. Out of millions. At that point, the law should have declared the margin too narrow to declare a victor and the next step in the process should kick in - be it a recount, a decision by the legislature, or even a coin flip.
I'm not asking for hocus pocus. I'm asking for a simple measure that honestly describes how accurate the process is. If you don't meet that threshold, you go to Plan B. Recounts are a joke if the numbers you are talking about are lower than the precision of the voting/counting method. Just decide the election with a coin flip and stop the charade if that is your plan.
A statistical analysis would NOT have helped. Nothing you've suggested would have helped.
I can't see how you look at Florida and declare success. The US Supreme Court had to fly in and save the day (or ruin the day). The whole process was a colossal fuck-up.
A state with millions of votes was trying to come up with an "answer" when a few hundred ballots separated the winner and loser. You will never get an answer, because the answer is that it was a tie... as you said, the voters had no clear preference.
So go with Plan B, and that can be whatever you want it to be. Another election, refer the matter to the legislature, whatever. Just don't let it get to the point where the state abdicates it's duty to decide who won. There was no process that allowed for a tie, so the courts had to step in. That should be unacceptable, but we don't like to admit that an election can be a tie and we don't define the criteria for a tie.
So long as you define the margin of error prior to voting, I don't see how it moves the problem.
I also have no problem with recounts, so long as they are part of the predefined process and remain statistically valid.
Yes, we do have a great method to decide how to handle a tie - the problem is that we don't have a mechanism to declare an individual state election to be a tie.
If you are getting hung up on the semantics of me calling it margin of error, then call it something else. Precision works for me. The important thing is to accept that the system is not perfect and that every recount will also have errors (or will also be imprecise). At some point, all that will happen is random results as the number of votes delta between the candidates is below the precision of the voting method. You could recount or you could flip a coin - the result would be the same. At some point you have to declare the election a tie and move to a backup mechanism.
If the results are very close, you can do a recount and find out if you were right the first time.
If you allow that there might have been an error the first time, why are you so sure you'll get it right the second time?
I wasn't arguing against a paper ballot or a simplified ballot - they both seem like good ideas to me.
I'm arguing that the election should be declared a "tie" if the margin falls within (for instance) the 2% error. Then backup solutions could kick in. The frustrating part is that there exist backup systems, but today they never get triggered because no one codified what a spoiled election looks like.
In 2000 in the US, when the Florida machines counted a 400-vote margin, it was immediately clear that the election was a statistical tie. But instead of an orderly process, there were these pointless recounts until the Supreme Court finally decided to step in. Luckily, the US is stable enough that the government didn't have a major crisis. A president more like Nixon and less like Clinton could have easily made things much more exciting.
That's kind of my point - that there can in fact be an election that is a "tie"... a second type of result in addition to the current "winner" result. At that point, backup mechanisms could kick in so you don't get stuck with the Supreme Court jumping in to what should have been an orderly handoff to the legislature.
I was screaming back in 2000 that the whole quagmire in Florida could have been avoided if the Constitution had simply been followed.
The courts got involved because there was disagreement about whether or not the election was valid. In my opinion the election was quite obviously a statistical tie, and you are correct that the existing backup mechanism could have sorted it out. The problem is that there was no predefined criteria for an election that was simply too close to call.
No. I'm talking about having a clearly defined system for resolving an ambiguous result. In other words, not just hoping that the government holds together when the Supreme Court decides to step in and impose such a system after the fact.
So, yes - one possibility is to hold the election again - perhaps as they do in much of the world where they narrow the field to the two top candidates. Another is to allow the legislature to decide. Another is a coin flip. It doesn't matter - just so the process is clearly defined and considered legitimate.
And the election is not a statistical analysis of the votes of the population, it is the actual votes.
If you are telling me with a straight face that millions of ballots are counted with no mistakes, I have a bridge to sell you.
For an election to work the system must be held to a standard of infallibility and that all errors fall on the voter,
That's impossible. No system is infallible.
if it's found not to be the case and is significant to have possibly affected the outcome a re-election is called.
Which is exactly what I am calling for. A set of criteria, that if met, declares the result to be invalid. I'm just asking for the rules to be clearly defined in the beginning.
So the margin of error is factored in
No it isn't. You said so yourself - the system is presumed to be infallible.
The recount thing is cute, but I'm not talking about a recount, which will of course also have a margin of error.
There is a fundamental flaw in elections today: lack of consideration for "margin of error". In my opinion, margin of error should be calculated and any election which falls within the margin of error should either be held again or some sort of tie breaker should kick in.
Pretending that we can deduce the intention of every voter with zero errors is noble, naive, and ridiculous.