"Did Kerry Win the Election?" Analysis

[russ_watters]

What is "the president isn't elected via exit poll" supposed to mean anyway? Who suggested that was the case? When different forms of data gathering produce inconsistent results, rational people investigate the reason.

It is implied in the link that if the exit polls show something different from the acutal polls, the polls must be wrong:

...if exit polls from various states use the same scientific methodology, then the likelihood of election results being significantly different than exit polls results in half a dozen swing states is very very low.

And "looking for justifications to challenge future elections"? Now there's an objective assessment for you. It seems it would not occur to you that having transparently verifiable procedures is intended to prevent challenges to elections by making flaws in the system more readily apparent and correctable.

And you ask me if I read the link? This is fairly clear:

...so that we can develop and test the efficacy of a system to put in place by 2006 to pinpoint counties or even precincts which warrant recounts.

That's clearly looking for a basis to have recounts, not prevent them. Its looking for a way to quickly evaluate where they think they can do recounts - to help the next Democratic canditate decide faster, where to send his lawyers-on-gulfstreams (Kerry's most important innovation in campaigning).

But if you think that miscounts, whether from software bugs or human error or fraud, are ok, then go ahead, argue against redundant verification methods.

That link says nothing at all about redundant verification methods. It isn't about fixing the election process, its about exploiting its flaws.

What part of the methodology is "suspect"? How does it "[ignore] the people who decided the election"? All that's been done is to point out a statistical pattern.

Right: a statistical analysis of democratic and republican voting patterns. This election was not decided by registered democrats or republicans (they rarely ever are): it was decided by independents.

 

[Gokul43201]

http://ustogether.org/Florida_Election.htm

Either this is really shocking...or I'm not following.

Consider little Calhoun County : Of the total of 8,350 registered voters, 82.4% were registered Democrats. Less than 1,000 were registered Republican. Now the total number of votes from this county was 5,961, out of which 3,780 went to Bush ! Assuming all those registered Rep or Indep (1,470 people) voted for Bush this still needs about half the registered Democrats who voted to have switched loyalty and voted for Bush. That can't be !

 

[franznietzsche]

Retracted post-momentary rage.

 

[BobG]

Physical location is probably more significant than the voting machine used. Every county in the panhandle East of Walton County had incredibly high results for Bush and accounts for most of the most drastic surprises on the list.

 

[plover]

Either this is really shocking...or I'm not following.

Consider little Calhoun County : Of the total of 8,350 registered voters, 82.4% were registered Democrats. Less than 1,000 were registered Republican. Now the total number of votes from this county was 5,961, out of which 3,780 went to Bush ! Assuming all those registered Rep or Indep (1,470 people) voted for Bush this still needs about half the registered Democrats who voted to have switched loyalty and voted for Bush. That can't be !

This is actually not the anomaly in the data. The low population counties in Florida are rural counties populated by conservative folks who are registered Democrat because of long standing patterns in Southern politics. I looked at the data for one county over time, and their choice for President seems most often to be the candidate who might be defined in some fashion as "more Southern" rather than that of a specific party, though it's not completely cut and dried. There is, however, no reason to believe that these counties would not have voted for Bush as the results indicate (indeed if these results were, in fact, as glaringly off as they appear at first glance, I assume that residents of those counties would have noticed and raised a fuss).

The recent Presidential races for Taylor County have gone as follows:

2004  Kerry    35.5%  Bush      63.7%                  Other 0.8%
2000  Gore     38.9%  Bush      59.6%  Nader     0.9%  Other 0.6%
1996  Clinton  44.8%  Dole      39.9%  Perot    14.3%  Other 1.1%
1992  Clinton  35.6%  Bush      37.3%  Perot    26.7%  Other 0.3%
1988  Dukakis  30.0%  Bush      69.1%                  Other 0.9%
1984  Mondale  30.0%  Reagan    70.0%                  Other 0.0%
1980  Carter   50.5%  Reagan    47.3%  Anderson  1.3%  Other 0.9%
1976  Carter   62.3%  Ford      36.7%                  Other 1.0%
1972  McGovern 15.5%  Nixon     84.5%                  Other 0.0%
1968  Humphrey 18.6%  Nixon     15.7%  Wallace  65.7%  Other 0.0%
1964  Johnson  39.1%  Goldwater 60.9%  Unpledged 0.0%  Other 0.0%

Note the winners especially in '64, '68, and '80.

The problems in the data show up in the fact that counties with urban centers (e.g. Leon and Duval—containing Tallahassee and Jacksonville respectively) indicate cross party voting in ways that don't correlate well with other urban centers. Unfortunately some articles have started quoting these statistics indiscriminately, without waiting for a sufficently rigorous analysis to be made. If there is anything to the statistical anomalies that ends up being explicable only by flaws in particular voting methods, then there will be something to be concerned about.

 

[franznietzsche]

Their just angry because they lost and they refuse to admit that it might have been because they were actually less popular. So much for Democracy. Its unconceivable to them that the majority might disagree with them, unimaginable.

Its a quite common perspective out here in california. Most people out here think conservatives are about 35% NATION WIDE, like they are locally. These people just can't concieve that maybe some people really wanted Bush to be reelected.

People disagree with me? NEVER...

 

[selfAdjoint]

Physical location is probably more significant than the voting machine used. Every county in the panhandle East of Walton County had incredibly high results for Bush and accounts for most of the most drastic surprises on the list.

The voting machine link was highly significant (p < .001) even when other factors were controlled for. RTFP.

 

[plover]

It is implied in the link that if the exit polls show something different from the acutal polls, the polls must be wrong:

...if exit polls from various states use the same scientific methodology, then the likelihood of election results being significantly different than exit polls results in half a dozen swing states is very very low.

And the sentence you quote here oversteps the bounds of statistical inference, um, how?

The woman who set up this page appears to have written it as if everyone reading it was going to understand that it was an incomplete analysis and not jump to conclusions, which was, of course, absurdly naïve. Almost every comment on this I've seen from either side (and I'd been following them for a couple of days before selfAdjoint's post) has done exactly the opposite.

And you ask me if I read the link? This is fairly clear:

...so that we can develop and test the efficacy of a system to put in place by 2006 to pinpoint counties or even precincts which warrant recounts.

That's clearly looking for a basis to have recounts, not prevent them. Its looking for a way to quickly evaluate where they think they can do recounts - to help the next Democratic canditate decide faster, where to send his lawyers-on-gulfstreams (Kerry's most important innovation in campaigning).

So, recounts are de facto bad? Or having accurate staistical methods which show where a recount might be useful is more untrustworthy than guessing? Or if the kinds of procedures Dopp is describing were available, both sides wouldn't be using them? You seem to have pretty much wandered off into partisan paranoia land here.

The more transparent the system is the better it is for both sides. Statistical measures are one method of increasing transparency. The victors (unless, of course, they are stealing elections... ) should want the vote results to be transparent so their legitimacy isn't challenged. Or do you enjoy living in a country with a snarlingly bitter partisan divide?

In my opinion, it would be extremely preferable if there were voting procedures in place that I trusted, but currently we have error-prone electronic voting machines with completely half-assed security. So as it stands I want to see this analysis done right—if there is something shady to find I want it found, and if there isn't anything beyond some counter intuitive patterns, I want that to be made clear so people don't sit around whinging and making up idiot conspiracy theories. Uncertainty and suspicion are a much worse hell than a fair and obvious loss.

That link says nothing at all about redundant verification methods. It isn't about fixing the election process, its about exploiting its flaws.

If I said that you didn't want voting procedures to be transparent so that Republicans could continue stealing elections, you'd say I was raving. I fail to see how the above theory is any more sensible.

Right: a statistical analysis of democratic and republican voting patterns. This election was not decided by registered democrats or republicans (they rarely ever are): it was decided by independents.

What does independents deciding the election have to do with an analysis of anomalies in cross party voting patterns?

 

[plover]

Computer security expert Bruce Schneier on electronic voting machines.

The auditing that is conducted on slot machine software in the U.S. is significantly more meticulous than what is done to voting software. The development process for mission-critical airplane software makes voting software look like a slapdash affair. If we care about the integrity of our elections, this has to change.

Proponents of DREs often point to successful elections as “proof” that the systems work. That completely misses the point. The fear is that errors in the software -- either accidental or deliberately introduced -- can undetectably alter the final tallies. An election without any detected problems is no more a proof the system is reliable and secure than a night that no one broke into your house is proof that your door locks work. Maybe no one tried, or maybe someone tried and succeeded...and you don’t know it.

Even if we get the technology right, we still won’t be done. If the goal of a voting system is to accurately translate voter intent into a final tally, the voting machine is only one part of the overall system. In the 2004 U.S. election, problems with voter registration, untrained poll workers, ballot design, and procedures for handling problems resulted in far more votes not being counted than problems with the technology. But if we’re going to spend money on new voting technology, it makes sense to spend it on technology that makes the problem easier instead of harder.