[NWPT48]I can’t tell you how disappointed I was to wake up this morning and find the TVW feed still working. Ah well, I suppose I better listen, and update this post as warranted. As always, The Seattle Times’ David Postman will also be providing regular trial updates.
The big news today should come out of the Frye hearing, in which the Court will hear the scientific basis for allowing the GOP’s proportional deduction methodology.
Here comes the exciting part…
That is, if you call testimony from statisticians exciting. The Dems have finished crossing Clark Benson, and the R’s main expert witness, Jonathan Katz has been called to the stand, so finally, the Frye hearing begins. Katz nervously commented to the judge that he’s never done this before, and didn’t know how this was going to work. To which Judge Bridges amusingly responded, “Well, I think Mr. Burman is going to work you over.”
Sure hope so. Stay tuned….
Burman works over Katz (9:54 am)
Polite, but firm. (Okay, now he just made a comment that wasn’t so polite, demanding that Katz look at him and answer the questions, rather than looking furtively to Rossi’s legal team for help.) Katz refuses to vouch for the reliability of the data set, and cannot say whether it was a scientifically designed or random sample.
Witness needs more water (10:15 am)
Katz has asked the court for more water. Burman’s thirst producing questioning focused both on the reliability of the data, and accepted scientific standards. Burman read from several sources that Katz vouched for, which laid out the principle that even the best available methods of ecological inference can be “wildly inaccurate,” and that in political science it is unreliable to infer from aggregate data. Katz admits that his statistical model relies on several assumptions, including “neighborhood homogeneity” and the reliability of the data.
Burman has finished with Katz (for the moment.) The Democrat’s statistical expert, Christopher Adolf, will testify after the morning recess.
Mmm… herring (10:33 am)
I took advantage of the break to surf around and see what others are writing, and I thought I’d point out this gem from our good friend Stefan over at (u)SP:
Burman brings up the red herring of the “ecological fallacy”.
That’s right, when you don’t have a rebuttal, just ridicule and dismiss the opposing side’s argument. Sheesh… sometimes I think Stefan practices his rhetorical technique by reading instructions from a how-to book. I Googled “ecological fallacy” and got 15,800 hits. Research it for yourself.
Good social science starts with good data (11:01 am)
That is the testimony of Prof. Adolph. Under Burman’s questioning, he is really tearing apart the GOP’s statistical analysis. I’m too busy listening to the testimony to comment in detail, but much of what Adolph testified to was covered in DJ’s excellent analysis of the his declaration here.
Adolph is now being cross-examined, and still sounds pretty confident.
Why not ask them? (11:21 am)
The Republican’s main refutation of Prof. Adolph’s testimony is that he is a non-tenured, former graduate assistant of the Prof. Katz, who he is now criticizing. Under cross-examination, Adolph suggested that an accurate way for determining the votes of felons might be to conduct a survey. When asked how one would possibly conduct such a survey, he bluntly proposed “since we have a list of illegal voters, I suggest asking them.”
They’re done with Adolf, and moving on to Prof. Mark Handcock.
GOP whimsy (12:51 pm)
Under questioning from Burman, Prof. Handcock stated that the “Principle of Insufficient Reason” was regarded in the statistical community as a “whimsical” notion. What wasn’t whimsical was the way the GOP attorneys tried to discredit Handcock, by showing that he doesn’t write about election data. Not surprising, considering he’s a statistician, not a political scientist. I know they’re just doing what lawyers do, but still, that doesn’t mean they don’t come across as assholes.
The main battle was over whether the GOP’s statistical methods could be applied to available evidence. The Democratic experts argued that the data was neither a scientific sample or represented the “universe” of illegal ballots. The Republicans’ response was that if the court accepted these ballots as the universe, then the methodology would be good. Handcock made a distinction between law and science.
The main point, is that while one could perform such an analysis on the available data, it still wouldn’t be scientific, no matter what the court ruled. And the R’s fight back that, well… the Dems don’t offer an alternative. But bad science is bad science.
This whole exchange is kind of like my eight-year-old getting pissy because she wants a bagel for breakfast and I don’t have one. I can’t give her what I don’t have. Likewise statistics simply doesn’t offer the court good science to accurately estimate for whom felons voted.
Court is adjourned until 1:30, and I’m taking my puppy for a walk.
John Carlson Show again at 3:15 (1:34 pm)
I just got a call, and John has invited Stefan and me back on the show today and tomorrow at 3:15 pm. I assume we’ll be talking about herring.
Court’s back in session. More later….
How the Judge WILL rule on Frye (2:03 pm)
I predict the Judge will adopt the Sec. of State’s position, that the evidence is admissible, but that it may not meet the “clear and convincing” standard of proof. This would be consistent with his rulings all along, that he allows nearly all the evidence in, and will eventually make his ruling on the evidentiary findings, not a point of law. This will make it very difficult to appeal.
How the Judge DID rule on Frye (2:16 pm)
Well, um… he didn’t.
Judge Bridges enunciated his “desire to make sure that as much evidence is available as possible” to the Supreme Court, which as I previously stated, has been his wont all along. And so he deserved to reserve ruling on Frye until later, and though it may not meet the letter of Frye, permit the parties to go ahead and present their cases. He did, however, address GOP attorney Mark Braden directly on the subject: “I have concerns.”
So… I suppose my prediction was somewhat accurate. The evidence will be presented, but he hasn’t necessarily accepted the science (indeed, he telegraphed that he might not.) That will be decided later, as well as the issue as to whether it meets the standard of proof.
And that bit of non-news was the biggest piece of news all week.
What David said (4:45 pm)
The court adjourned early today, and I have no idea why. After 45 minutes on the Carlson show and fielding a couple phone calls, I came back to find it all over. Not that I probably missed anything, as they were just taking testimony from expert witnesses who had previously been deposed. But if you’re really curious, I’d suggest you just read the latest updates from David Postman.
Goldy,
You don’t have to update us on the trial. After all, you are the entertaining one, not the intelligent one. Remember?
pbj… yeah but I’m apparently too stupid to realize that.
on that note, Clark spells his last name with 2 e’s. Bensen. And don’t feel bad about it, TVW got it wrong, Archerhouse at Kos got it wrong, and about 2/3 of the time I wrote it in my Day 3 recap, I got it wrong too.
Katz had way too much caffeine this morning.
The attorneys may have, as well, but I’m listening for more signs of sleep-deprivation like yesterday’s slurring of words (one of the R’s, at least).
Oh boy, this was a great exchange:
He had Bensen agree that there is some uncertainty to the data. And then he got to the point.
“You’d be really offended,” Burman said to Bensen, “if I accused you of a fraud because you are not absolutely certain of the accuracy” of the records.
That’s what Democrats say Republicans have done with King County election officials — take uncertainty and errors and call it fraud.
As I listen to you poor guys covering the trial, I am really happy that I have a job to go to!
David T
Ah, so Shark wants his readers to believe that mainstream statistical thought, backed by over a century of analysis and theory, is a “red herring”.
Just goes to show what happens when someone as clueless as Stefan tries to pontificate about something where he’s waaaay out of his league … I’m imagining my 11-year-old Little Leaguer nephew standing in the batter’s box against a major league pitcher (sorry, kid, about vaguely equating you with Sharkansky).
Unsound, indeed.
At least Katz seems to be a AA or AAA journeyman trying to catch on in the bigs … beyond his skill level, but (unlike Stefan) at least he has some of the basic skills.
Goldy or anyone else who is watching the trial,
KEY QUESTION: Did the DEM lawyers hammer Katz on the “homogeneity” assumpton? I.e. the assumption that the precinct sample was representative of an illegal voter from that precinct.
I think this will be the key issue for Bridges in deciding whether to admit the proportional reduction evidence.
However, he might admit it just to be safe, and then find it isn’t clear and convincing proof.
chew2 @ 8
” Did the DEM lawyers hammer Katz on the “homogeneity” assumpton? I.e. the assumption that the precinct sample was representative of an illegal voter from that precinct.”
Indeed, they did. Unfortunately, I lost the feed twice for big chunks of his testimony, so I am not sure about the extent of their “hammering.”
To a large extent, the questioning of both expert witnesses moves fluidly between methodological issues and data issues.
The two issues are strongly linked, of course, because the best methods would either explicitly account for (or partially account for) the heterogeneity in the aggregatedata, or taking into account the limits of the heterogeneity (i.e. a proper ecological inference method).
Goldy,
I have to say, for a newly minted Ph.D., Professor Adolph was simply outstanding. He communicated clearly, confidently, and was unflappable.
I was amused by the little jab about the former graduate research assistant testifying against the professor. Apparently the Petitioners feel being a former graduate student of Gary King (perhaps the most esteemed Political Scientist in America) and Jonathan Katz (who Burman touted as one of the most esteemed Political Scientists in America) somehow makes his testimony less valid? I wonder who Judge Bridges will find more credible based on their statements and testimony :-)
dj @9
As noted before, we disagree about whether the biased or incomplete sampling of their data by the GOP is a flaw. As I said, the GOP is only using proportional reduction to show how the IDENTIFIED illegal voters may have voted. It’s my view that legally, they do not need to prove their sample was unbiased to carry their burden of proof. There is a (weak) legal argument to the contrary, but I don’t think Judge Bridges will buy it. So the key issue is homogeneity.
BTW what did Katz say to justify his assumption of homogeneity? I don’t recall much discussion of it in the deposition (because the DEM lawyer flubbed) or in his report (where he purposely glossed over this issue).
An assistant professor is the best “expert” the Democrats could find? What, the TA wasn’t available? Goldy, did they ask you? Even though you reportedly aren’t the “intelligent” one, somehow I don’t think that matters to the Democrat attorneys.
Comment on 12
There you go again, labeling people — in a lame effort to distract from the actual data.
chew2 @ 11
“As noted before, we disagree about whether the biased or incomplete sampling of their data by the GOP is a flaw. As I said, the GOP is only using proportional reduction to show how the IDENTIFIED illegal voters may have voted. It’s my view that legally, they do not need to prove their sample was unbiased to carry their burden of proof. There is a (weak) legal argument to the contrary, but I don’t think Judge Bridges will buy it. So the key issue is homogeneity.”
I understand your argument and the merits of it. The counter argument is that because each invalid voter’s vote is not known, a scientific method must be used to estimate (1) the expected winner of the election and (2) the certainty with which that can be established.
If the courts adopt a scientific method, then they should do so in a way that is scientifically valid. Without a random sample of invalid voters, or a complete census of invalid voters, or some way of adjusting for underenumeration of invalid voters, neither 1 nor 2 can be determined using acceptable science.
In other words, if we knew how each invalid voter voted, the courts could potentially rule using only the cherry picked votes if they wished—that is a purely legal question. But, if you invoke science because science can indirectly answer the question, you must stick to valid science to ensure the answer obtained by the indirect method actually “worked”.
“BTW what did Katz say to justify his assumption of homogeneity? I don’t recall much discussion of it in the deposition (because the DEM lawyer flubbed) or in his report (where he purposely glossed over this issue).”
Perhaps someone who caught a continuous version of the testimony can comment. The pieces of testimony I caught make it difficult to characterize his answer.
dj
The counter argument is that because each invalid voter’s vote is not known, a scientific method must be used to estimate (1) the expected winner of the election and (2) the certainty with which that can be established.
No, what must be proved is a LEGAL QUESTION not a scientific one. I don’t believe the GOP has the legal burden to prove the winner of the election taking into account all illegal votes. They carry their legal burden of proof if they identify enough illegal votes to change the outcome. It’s then up to the Dem’s to come up with additional proof to counter that initial showing, including the fact that there are other illegal votes out there.
dj,
Here’s a question that should have been asked at the Katz depo. If your assumption of homogeneity is flawed, what does your analysis tell us and what is the range of error in it? How do you know?
That’s right, when you don’t have a rebuttal, just ridicule and dismiss the opposing side’s argument. -by Goldy, 05/26/2005, 9:23 AM
Did you actually write that with a straight face, or is that a bit of your hidden irony?
Fun to see you guys dissing Shark on statistics. I’ll bet he’s the only one posting that actually understands them. (certainly not the elections officials…ouch)
pbj @ 12
Got any specific criticisms of Professor Adolph, or are you just blowing smoke out your ass?
“An assistant professor is the best “expert” the Democrats could find?”
Let me share some information about Adolph. Two years ago, the Center for Statistics and the Social Sciences at UW conducted a search for someone who is both a statistician and a social scientist at the Assistant or Associate Professor level. The person could have a degree in any social science (or statistics). The successful candidate would be affiliated with the most appropriate social science department on the UW campus. I should point out that the UW is in the top 5 Universities in mathematical and statistical sciences, and, if they broke this out, the UW is probably the best university in the world for statistics in social sciences.
There were 160 applicants for the position, and Christopher Adolph was one of a very small number of individuals invited to interview–he was the only person to get a job offer. In other words, he competed with the best candidates from all social science fields and got the prize. This makes Professor Adolph a hot commodity in the world of social science and statistics. He has a 2005 PhD from Harvard where he worked with Gary King and Katz. Aside from hiring Gary King himself, it is hard to imagine finding a better expert witness.
chew2 @ 15,
“No, what must be proved is a LEGAL QUESTION not a scientific one. I don’t believe the GOP has the legal burden to prove the winner of the election taking into account all illegal votes.”
I believe you missed my point. Because there is not a straightforward way to address the legal question, the courts wish to make a legal decision informed by a scientific analysis.
The problem is that the scientific analysis itself has requirements to ensure scientific validity. If those requirements are not met, the science cannot be claimed as valid.
The courts may choose, if they wish, to use the science and then to ignore the fact that the science is not valid (and they certainly may do so here). The reason for a Frye hearing, I believe, is to determine whether the courts can rely on science to provide insight into a legal question. If the science is not valid by scientific standards, my understanding is that the courts are supposed to disallow it. (I’m not a lawyer, though, so insights from lawyers would be most helpful for this discussion.)
So, in short, the “random sample” is needed for valid deployment of the science, whether or not it is a requirement for the legal case. Translating the science (whether valid or not) into a decision is a legal question.
Well, Sharkansky does have a degree in Mathematics, but he clearly didn’t specialize in statistics. I’ve just seen him make far too many idiotic statements with respect to stastics in general for me to believe he specialized in that area of mathematics all that much. Remember, it doesn’t matter how many letters you have after your name, if you say something stupid, it’s still stupid.
More importantly, however, Sharkansky certainly has very little background in the proper use of stastics in social sciences. For Sharkansky to use a “red herring” label with respect to the ecological fallacy really reveals way more about his lack of knowledge than it does about the issues at hand.
You know, what is really sad about all of this is that the ideology of the dems, and quite frankly just about everyone here, is that preserving Gregoire is more important than knowing who actually won, or trying to do anything to resolve the problem. Every single screwup in this election is discounted by everyone from Sims to Goldy as par for the course, but if the shoe was on the other foot and Rossi won the last count, you would be spinning every error as proof the election was stolen. You have no integrity.
chew2 @ 16
“Here’s a question that should have been asked at the Katz depo. If your assumption of homogeneity is flawed, what does your analysis tell us and what is the range of error in it? How do you know?”
I believe this was asked because Adolph later said that Katz discussed exchangeability (which is the same issue as an assumption of homogeneity). I missed that, but I did hear Katz invoke something like “my analysis was conditional on the data” when I got the feed back, so I am guessing the exchangeability discussion came up just before that.
dj @ 20,
We keep repeating the same arguments. I don’t know if we are blockheads or just stubborn.
Try this question.
Let’s say you have identified 2 illegal felon votes in a precinct. Can you use proportional reduction as a means of estimating who they might have voted for. If so, why and with what error? If not, why not?
Itsasquak @ 22:
Actually, there’s absolutely NO WAY of determining who actually won the election.
None.
The margin of victory was well within the margin of error of even the best election systems. It’s simply not possible to hold an ANONYMOUS election with 2.5 million votes and expect 100% accuracy.
If your vote wasn’t anonymous, then this could be done, but as long as the element of anonymity is entered into the equation, then there is too much information lost to ensure accuracy.
Every time I hear people talking about how shoddy this election is, I shake my head in wonder. IMHO, this election was a MODEL of accuracy that should be admired across the world – yes, there were mistakes made, but the level of inaccuracy in this election is extraordinarily low.
Ralph Reed, and the rest of the people involved in this election should be justifiably proud of the work they did, but all they’re getting is villified in the press/blogosphere.
Reply to 22
You’re new here, aren’t you? I (and several others) have repeatedly posted on HorsesAss that if the evidence shows Rossi won, then he is our governor.
You seem to be parroting the right-wing refrain that Gregoire is not a legitimate governor because the election was flawed. If you adopt that reasoning, then Bush is not a legitimate president, and hardly any elected official occupies the office legitimately, because no election is perfect.
We have a legal process to challenge elections. Rossi has the burden of proving the errors changed the outcome. The problem is that he and the GOP have gamed this election ever since Nov. 2, and they are gaming the lawsuit as well. Instead of offering proof that Rossi was robbed, they’re offering a statistical methodology that doesn’t satisfy the scientific standards of the statistics profession for scientific validity. If that’s all he’s got, then I would say, “It sure looks to me like Gregoire, not Rossi, actually won.”
Larry @22 The first two counts WERE model elections as the desired result was reported. It was only when Rossi was found to have lost that any problems arise.
Sorry Larry, my bullshit meter just broke. Any yayhoo that talks about this election being a model of accuracy is a democrat partisan hack, and that is exactly what I am talking about. You just proved that you have no integrity.
Undervotes, overvotes, felon votes, dead people votes … if the election was decided by anything less than the sum total of the votes that appear to exist in those catagories combined, then we have an election that is in doubt because of the margin of error.
I have not heard one single proposal from the democrats about what should actually be done about this. The basic premise I get on here is that the Republicans should quit bitching and accept the way things came out, regardless of the doubts raised by the OV, UV, FV and DV. My point is, no … I am a voter in my state, and I want a legitimatly elected governor, and that means I dont want there to be any doubt as to whether my vote counted or not.
If your vote disappeared, you are disenfranchised. If votes were fabricated, we are all disenfranchised. If people who should not have voted did vote, then we are all disenfranchised. The only time we should say the election is over is when the results would be unchanged even if all of the sources of potential disenfranchisment went against the winner.
WHAT, if the shoe was on the other foot and Gregoire lost the last recount, would be YOUR proposal for actually making this election whole to the point where you would shut up and accept it.
Itsasquak @ 22
“You know, what is really sad about all of this is that the ideology of the dems, and quite frankly just about everyone here, is that preserving Gregoire is more important than knowing who actually won, or trying to do anything to resolve the problem.”
That is not a “Dem thing” or even a liberal thing. The legal system is preserving Gregoire’s win. The laws of the state require the GOP to prove that Rossi won the election. This is true regardless of which party claims injury.
“Every single screwup in this election is discounted by everyone from Sims to Goldy as par for the course, but if the shoe was on the other foot and Rossi won the last count, you would be spinning every error as proof the election was stolen.”
Perhaps the Dems would challenge (I’ll exclude myself here because I have been politically unaffiliated for my whole life). The fact of the matter is that most of the errors that occur are ordinary (e.g. felons and dead voters are found in every statewide election that is scrutinized). Most of the errors are minor and constitute a tiny fraction of the ballots cast. Finally, none of the errors have conclusively shown that Rossi really won.
I can imagine errors for which I would gladly see the election conceded to Rossi. For example, suppose we found that a software glitch in King County Accuvote machines registered a Gregoire vote every 16 ballots, regardless of who the voter actually voted for. (Of course they would simply re-run the ballots through after fixing the software, but ignore this possibility for this example).
A simple proportional analysis of the KC vote, in that case, would be completely appropriate. We would know that a random sample of 1 in 16 votes was probabilistically switched to Gregoire. We could easily get an unbiased estimate of the number of votes for each candidate and a confidence interval around that estimate. Almost certainly, Rossi would have won in this hypothetical case, and I think many Dems would accept that finding. (Some would be pissed, but they would accept it).
Likewise, we found documentation that Dean Logan personally communicated with poll workers in many precincts and told them to stuff one or two ballots through the machines with a Gregoire vote, there would be liberal outrage. But, the GOP legal team has not found anything even close to this.
dj@19,
Adoph has no degrees in mathematics or statistics. His degrees are in Political Science. And he is an assistant professor, not a tenured professor.
As for all you bluster about the UW being the best in the nation, please provide some documentation (sources) to back that up.
UW also thinks they have the best Computer Science program in the nation too, but they cannot even qualify for CSAB (CS people know what that is).
Larry @ 25
Ralph Reed? Probably Sam Reed. But if Goldy ever wants to write a piece vilifying Ralph Reed, there’s a lot of good material out there.
I wonder if the FCC will slap one of its $500,000 indecency fines on right-wing radio host/hatemonger Glenn Beck for threatening to kill Michael Moore on the air? http://mediamatters.org/items/200505180008
Naaaaah … I’ll bet Bush’s FCC gives Beck a free pass.
Rick@21,
What is your CV that makes you such an expert on statistics?
chew2 @ 24
“Let’s say you have identified 2 illegal felon votes in a precinct. Can you use proportional reduction as a means of estimating who they might have voted for. If so, why and with what error? If not, why not?”
Under some conditions you can “reduce” the votes statistically.
One important condition is that you must assume that the felons votes are “exchangible” with all other voters with respect to their voter preference. That is, the probability of them voting for Gregoire is independent of the fact that they are a felon.
Of course, you must know the number of ballots cast for each candidate by the entire precinct.
If those assumptions hold, you can do a reduction by drawing the fraction of votes to remove for each candidate from a trinomial with the candidate proportions. The error of the method can be determined using the formula for the standard error of a trinomial. Alternatively, you can simulate the process over and over again a large number of times and look at the average outcome and empirical error rate.
When one examines all the accuracies required to launch a man 384,400 kilometers to the moon and bring him back safely such as we did in 1969, one wonders why King County cannot accuractely count votes in an earthbound election.
Equally astonishing is how Chelan County, using the exact same software, managed to have an accurate count. No need for falsifying election reports in that county.
But listening to all the liberal partisan Democrats you’d think that counting a million votes was somehow equivalent to splitting the atom.
dj@29,
“Likewise, we found documentation that Dean Logan personally communicated with poll workers in many precincts and told them to stuff one or two ballots through the machines with a Gregoire vote”
– dj
Freudian slip?
Or shall we say Fruadian slip?
Comment on 29
Nindid was being facetious, you idiot. Since you’re too dumb to figure it out, I’ll help you. What he really said was:
“[The Republicans thought] [t]he first two counts WERE model elections [because] the desired result was reported. It was only [after the hand recount] when Rossi was found to have lost that [Republicans had] any problems [with how the votes were counted].”
Ooops, that should be Comment on 28, not 29. I was replying to Itsasquak.
Reply to pbj @ 36
He was giving an example, dope! Once again you have taken something someone said out of context and twisted the poster’s meaning beyond all recognition.
Addendum to Reply to 36
And you can’t even spell “fraud” correctly …
dj
Thanks for the answer. This is what the GOP expert is proposing to do.
“One important condition is that you must assume that the felons votes are “exchangible” with all other voters with respect to their voter preference. That is, the probability of them voting for Gregoire is independent of the fact that they are a felon.”
Do we have enough info to suggest that this is not the case? E.g. that felons here are 75% male.
“The error of the method can be determined using the formula for the standard error of a trinomial.”
And if so, what does that do to your error estimates. This will be key to the DEM’s case, since they have lost the Frye hearing. My belief is that Katz’s error estimates were always much too low, but it’s incumbent on the DEMs to show why and to suggest some error bounds of their own. Otherwise they may be in deep shit!
pbj/harry/joseph your partisan echo-chamber ignorance is breathtaking.
Maybe the elections systems in this state will approach the kind of perfection you wingers pine for when they have the kind of budgets NASA did in the sixties (adjusted for inflation) and even then a few mistakes will be made like oh astronauts getting killed during training and almost getting killed during a mission.
Chelan County is a straw man, pbj. 20 some thousand voters and a handful of polling places. Nice try.
By the way, fraud is spelled F R A U D so your joke fell flat. Also dj’s “slip” was dropping an “if” in one of his sentences, I believe.
I still don’t get it. Is there no way to find or prove fraud in elections in this state, short of total and absolute confession? Imagine the same laws to prove someone is a Mafia don?
Evidence points to something amiss, but the recount process may not allow that to be considered. So we are on the verge of finding gobs or errors or fraud, and zip way prosecute the theft?
Quick question – my feed has been sketchy, and I’ve missed some details. Is the current analysis being described (put forth by Republicans) based only on the Republicans’ alleged fraudulent voters? Or did they re-do their analysis to account for the Democrats’ list? I know dj has done some analyses showing how unlikely it would be that even this dubious method being put forth would end up working in the Republicans’ favor if the Democrats’ list is included. So which one is currently being discussed? Thanks…
righton @ 18
there are three people just writing comments here who have a demonstratively better command of social science statistics: dj, chew2, and (humbly) myself. We’ve caught Sharkansky REPEATEDLY in statistical theory errors and analyses.
dj @19
how about UM’s ISR? Mike Traugott’s got some pretty sharp SS minds there, although I suppose narrowly speaking on statistics, you may have a good argument.
On the first problem with ecological inference (I think we all three agree that the simpler failing is non-homogeneity): I see both points have merit. In the larger trial, I would agree that all Rossi has to do is present illegal votes, such that they show he won. He does NOT have to suggest that they represent ALL illegal votes; unknown illegal votes are presumed valid in court.
However, in the context of a Frye hearing, it may be true that if the model itself fails, then the use of the model must not be allowed. And in that context, it’s true that their analysis is suspect because we know they don’t have the full universe of illegal votes. So before it ever got to the point that Rossi could show his votes turned the election, he could have his method disallowed.
In any case, even if the method is allowed, once the Democrats present their own illegal votes, it brings strong counter to the argument that Rossi’s changed the election–since it can be shown that his presentation does not represent a fair collection of testable votes.
chew2 @ 41
I swear I heard Adolph report that (having actually tested for it), the error bound was Gregoire 0% to Gregoire 100%.
righton @ 43
If there are simply errors, the obvious assumption is that absent known fraud, the errors hurt each candidate on a random basis. That is, any group of mishandled ballots could have any breakdown of votes for each candidate. Only through intentional attempts to actively SKEW the results, can it be deduced that the result was changed.
As for how to prove fraud–read the Foulkes v Hays decision.
pbj @ 30
“Adoph has no degrees in mathematics or statistics. His degrees are in Political Science. And he is an assistant professor, not a tenured professor.
Gee. . . the GOP lawyers were giving Professor Handcock shit because he wasn’t a political scientist. You criticize Adolph because he doesn’t have a degree in statistics. Go figure.
Adolph was hired as core faculty in the Center for Statistics in the Social Science (CSSS), and is an adjunct professor in the Statistics Department (one of the best in the country). The fact that he is not tenured is because he hasn’t worked at the University for 6 years, which is almost always necessary before one gets promoted and gets tenure. He is hired on a tenure track line, however.
“As for all you bluster about the UW being the best in the nation, please provide some documentation (sources) to back that up.”
Here, http://www.stat.ucla.edu/program/ranking.php , is a summary showing that U.W. ranked at both number 6 and number 9 in 1995 based on a National Research Council report. I understand that they have moved up to number 2 and 5, but I haven’t found a more recent report from this source. The multiple rankings are because Biostatistics Department is ranked separately from the Department of Statistics. The fact that one university has two top-10 departments is nothing short of astounding.
You can go to http://www.phds.org/rankings/ and search by various criteria for statistics. Here are some of the more important intellectual quality measures:
Faculty quality: UW is ranked twice at #6 and # 9
Number of publications per faculty member: UW is ranked at #2.
Citations per faculty member: UW is ranked at number #2 and #8
Percent faculty with research grants: UW is ranked #6
One more “ranking”: Adolph and Handcock are 2 of 6 core members of the CSSS (see http://www.csss.washington.edu/). The director of CSSS is the single most cited mathematical scientist in the world for the last 10 years. From the CSDE web site:
Adrian Raftery , Professor of Statistics and Sociology, and Director of the Interdisciplinary Center for Statistics and the Social Sciences at the University of Washington, was the most cited mathematician in the world for the period 1995-2005, according to the Institute for Scientific Information … A researcher’s impact is often measured by counting the number of times his or her work is cited in scientific research articles by other scientists. . . Professor Raftery’s tremendous accomplishment is a tribute to his influence in the field, and to the esteem of his peers.
Look at the rankings for yourself here: http://www.in-cites.com/top/2005/first05-math.html
Judge Bridges enunciated his “desire to make sure that as much evidence is available as possible” to the Supreme Court, which as I previously stated, has been his wont all along. And so he deserved to reserve ruling on Frye until later, and though it may not meet the letter of Frye, permit the parties to go ahead and present their cases. He did, however, address GOP attorney Mark Braden directly on the subject: “I have concerns.”
Wow. This judge really does want to hear it all. He wants to be interlocutory appeal proof and he is succeeding at this. He knows that if he makes any harsh ruling one way or another, the case will be delayed while an issue is appealed.
Yes, his decision will be ultimately appealed. However, he may make a number of rulings which could almost inoculate him from that as well. Almost.
pbj @ 36
d’oh!
Likewise, IF we found documentation that Dean Logan personally communicated with poll workers in many precincts and told them to stuff one or two ballots through the machines with a Gregoire vote, there would be liberal outrage.
Thanks for pointing that out!
chew2 @ 41
“Thanks for the answer. This is what the GOP expert is proposing to do.”
Not really. If that assumption holds within a precinct, the answer you get only provides a valid answer for a precinct. It does not give you a valid answer for the collection of all precincts in the state. To apply the answer at the statewide level, additional assumptions are needed. So that is one scientific problem for the GOP.
Beyone that, if the exchangability argument is false then the results cannot be considered valid.
“Do we have enough info to suggest that this is not the case? E.g. that felons here are 75% male. “
I am not sure what you are asking. Prof. Handcock did do a sex-adjusted analysis in his report.
“The error of the method can be determined using the formula for the standard error of a trinomial.”
“This will be key to the DEM’s case, since they have lost the Frye hearing.”
I don’t think they lost the Frye hearing. . . Judge Bridges said he would defer his decision.
“My belief is that Katz’s error estimates were always much too low, but it’s incumbent on the DEMs to show why and to suggest some error bounds of their own. Otherwise they may be in deep shit!”
His error estimates are fine if the only source of uncertainty in his analysis is binomial sampling error. In the binomial version of his statement (report 2?), for example, he properly has a variance in the final vote outcome that is 4 times the variance of a binomial, which accounts for the negative correlation between the scores as votes are removed from each candidate.
Where Katz’s analysis breaks down is that it does not account for uncertainty reflecting non-exchangeability of the felon voters. I previously posted some other, more minor, criticisms of Katz’s method as well (using a binomial, and using the normal approximation to the binomial), but these sources of error are far less important than the uncertainty of non-exchangeability.
Reply to 43
“Is there no way to find or prove fraud in elections in this state, short of total and absolute confession?”
For starters, there has to be fraud …
torridjoe @ 46
“I swear I heard Adolph report that (having actually tested for it), the error bound was Gregoire 0% to Gregoire 100%. “
He did. That adds considerable uncertainty into the estimates (as in: statistically, no outcome from 0% to 100% is improbable).
36K people in 80 precincts or 1 million people in 2600 precincts. I wonder which is easier to keep track of.
Reed detroys Stefan’s excel spreadsheet theory
Well. Someone had to tell SP they were in left field. I guess it had to be a republican.
Yesterday, as the WSRP daily summary reports
The lawyer for the Secretary of State objected that voter crediting information could not be used to prove anything about who voted or how many people voted
dj @52
“considerable uncertainty”–:) That’s a polite way of putting it, when the analysis suggests any breakdown is mathematical plausible.
we don’t claim sam as ours
torridjoe @ 45
“However, in the context of a Frye hearing, it may be true that if the model itself fails, then the use of the model must not be allowed. And in that context, it’s true that their analysis is suspect because we know they don’t have the full universe of illegal votes.”
I would only add to your point my mentioning that valid use of the method doesn’t necessary require the “full universe of illegal votes.” One alternative is that we could have a random sample of illegal votes. Another is that we could have a biased sample, but additional information on how to weight each subsample.
BTW, the early look on the Bolton cloture vote is that the filibuster will be SUCCESSFUL. Only about halfway through the voting, but no defections yet. Feinstein and Lieberman both vote no on cloture.
torridjoe @ 55
Indeed :-)
dj @ 57
quite right; didn’t mean to imply otherwise. However, without a knowledge of the universe it’s awful tough to know whether you have a good sample.
Landrieu and Nelson defect on cloture. Jeffords with, so they’ve got 2 more defectors to spare.
dj@49,
I think you were being more truthful the first time!
Skinny @ 44
“Is the current analysis being described (put forth by Republicans) based only on the Republicans’ alleged fraudulent voters? Or did they re-do their analysis to account for the Democrats’ list?”
They are mostly discussing the methods themselves. Neither Katz or Gill have done any analyses with the Dem list as far as I know. I think the final analysis will include some or all of the data sets. The GOP has identified 8 or 9 different batches of ballots they want tossed out. The Dems have 4 batches of ballots to toss. The analyses done by Adolph and Handcock have only included 5 of the GOP batches and 2 of the Dems batches.
CLOTURE DENIED ON BOLTON–FILIBUSTER SUCCESSFUL.
41 No votes and counting…
John@42,
Excuses excuses. Maybe you could stop spending $300,000 on space age toilets and use the money for the elections department.
It is amazing that companies each and every day account for millions of items in inventory and yet King County cannot even manage to count votes correctly. Maybe they ought to contact American Idol.
dj@47,
“Here, http://www.stat.ucla.edu/program/ranking.php , is a summary showing that U.W. ranked at both number 6 and number 9 in 1995 based on a National Research Council report.”
The source for those “rankings” date back to information as old as 1982. 23 year old data hardly indicates cutting edge statistics programs for 2005!
Source 3 for the survey: 1982 Survey of Graduate Faculty
Here are some of ranking categories:
93Q
Rank order of “scholarly quality of program faculty” trimmed means for 1993 NRC study. Average score on a scale of 0 to 5 with 5 representing “distinguished”.
93E
Rank order of “program effectiveness in educating research scholars and scientists” trimmed means for 1993 NRC study. Average score on a scale of 0 to 5 with 5 representing “extremely effective”.
82Q
Rank order of Faculty Quality Score for 1982 Study, based on the average score on a scale of 0 to 5 with 5 representing “distinguished”. Source: 1982, An Assessment of Research-Doctorate Programs in the United States.
None of that information is recent. As far as citations, having lots of whacko liberal profs cite other whacko liberal profs only makes one “distinguished” in the eyes of other liberals.
“You can go to http://www.phds.org/rankings/ and search by various criteria for statistics. Here are some of the more important intellectual quality measures:
Faculty quality: UW is ranked twice at #6 and # 9
Number of publications per faculty member: UW is ranked at #2.
Citations per faculty member: UW is ranked at number #2 and #8
Percent faculty with research grants: UW is ranked #6″
If you actually go look at the sources you cited, you will notice that the high rankings are for the BIO-Statistics, not the statistics program at the UW.
BZZZZT. Thanks for playing. Try again.
One more “ranking”: Adolph and Handcock are 2 of 6 core members of the CSSS (see http://www.csss.washington.edu/). The director of CSSS is the single most cited mathematical scientist in the world for the last 10 years. From the CSDE web site:
pbj @ 64
American Idol…hmmm…isn’t that the show where they had to REVOTE one week this season because they got the phone numbers wrong?
Good choice there.
“One more “ranking”: Adolph and Handcock are 2 of 6 core members of the CSSS (see http://www.csss.washington.edu/). The director of CSSS is the single most cited mathematical scientist in the world for the last 10 years. From the CSDE web site:”
A leftist think tank hires a leftist prof as a “core” member. Imagine that!
joe@67,
Key phrase: “they had a revote”. See when it goes wrong the first time, they rectify rather than deny. Unlike KC elections where they just deny deny deny.
Thanks for pointing that out!
Actually, American Idol is fraught with many other problems with the voting. The revote was only because it was a colassal error on the scale of printing names in the wrong order on a ballot. And these problems have been publicized for years.
Honestly, anyone who looks to American Idol as a standard for elections doesn’t deserve to be taken seriously.
pbj @ 69
can you point out the section of Washington state law that authorizes a re-vote?
And nice attempt to move the goalposts. You were citing AI as a model of accuracy, not one of model mess-cleaning-up. When I note that your analogy was laughably flawed, you changed the game.
pbj @ 66
“If you actually go look at the sources you cited, you will notice that the high rankings are for the BIO-Statistics, not the statistics program at the UW.”
No shit, Shirlock. But, I was discussing the UW’s standing, overall specifically I said:
“I should point out that the UW is in the top 5 Universities in mathematical and statistical sciences. . . .”
With two top 10 (previously) or top 5 (currently) statistics departments, this is hard to argue that UW is the place to be for statistics.
BZZZZT. Thanks for playing, dipshit.
Every single screwup in this election is discounted by everyone from Sims to Goldy as par for the course, but if the shoe was on the other foot and Rossi won the last count, you would be spinning every error as proof the election was stolen. You have no integrity.
In my imagination, you’re a hypocrite!! Yeah, take that, liberals!!
DJ @ 51
Jesus F*cking Christ!!! I’m thinking you are being the blockhead now.
You keep repeating the same damn thing.
You say:
“Not really. If that assumption holds within a precinct, the answer you get only provides a valid answer for a precinct. It does not give you a valid answer for the collection of all precincts in the state. To apply the answer at the statewide level, additional assumptions are needed. So that is one scientific problem for the GOP.”
We disagree about what the GOP has to prove. Look, I’m a lawyer and it’s my opionion that the GOP only has to prove it for the precincts they have selected. My question asked for a single precinct, and you said proportional reduction would work for that. You keep repeating that they have to show it for all the precincts. Why don’t you accept that we disagree about this.
My question is can we demonstrate that Katz’s assumption of homogeneity is incorrect, and if so what does that do to his estimates and his estimates of error.
You say: “Where Katz’s analysis breaks down is that it does not account for uncertainty reflecting non-exchangeability of the felon voters.”
My question is if you do account for that “uncertainty reflecting non-exchangeability” what can you say about Katz’s estimates and specifically about his error bounds. Can you quantify how much this increases the error. Or is all you can say is Katz’s study is flawed….or what?
TORRIDJOE @ 46
You say: “I swear I heard Adolph report that (having actually tested for it), the error bound was Gregoire 0% to Gregoire 100%. “
In his report, Adolph stated that the number of illegal votes in any precinct was way to small to estimate any sort of bounds for the illegal vote, using Kings ecological regression methods. So that the Gregoire vote could be 0 or 100% using King’s methods. I believe this is what he must have been referring to. I don’t believe he was specifically commenting on the error bounds in Katz’s study, using Katz’s methodology and homogeneity assumption. But I didn’t hear the testimony.
But that’s what the Dems have to show, how bad is the Katz estimate if his homogeneity assumption is flawed? Can they quantify that?
It is amazing that companies each and every day account for millions of items in inventory
No they don’t fool. Private businesses have a notoriously bad history of tracking their own stuff. That’s why so many of them fail! Sure government is no shining example but neither is the private sector.
I don’t know where you’ve been but I’ve worked in grocery and retail and inventories are laughably inaccurate. Ever hear of the word shrinkage?
Bzzzt. Next.
So, the Republicans’ statistical method cannot infer the gender of the alleged illegal voters, but they want to allow it to infer for whom the alleged illegal voters voted? Using their methods predicts that ~50% of the illegal voters are male (with a tight confidence interval), when in fact >70% are male. And this is the method that’s supposed to convince the court that it can clearly and convincingly say that (enough) more of the illegal voters cast their ballots for Gregoire? Seems pretty shaky to me.
Good one in kc.
no shrinkage in a no customer store
Comment on 56
What’s a matter righty, when Sam lined up with the other GOP celeb politicians to pose for a photo op and demand a revote, that wasn’t enough for ya? Oh, I almost forgot, he “certified” the election … yup, there’s gonna be payback for that …
Comment on 58
GOP Senators fell 4 votes short of breaking the Bolton filibuster. Another defeat for presidential aspirant Bill Frist. Maybe he shouldn’t have quit his day job.
Comment on 64
You want us to believe that companies with “millions of items” in their inventories never lose anything? Bullshit.
skinny @ 76
“So, the Republicans’ statistical method cannot infer the gender of the alleged illegal voters, but they want to allow it to infer for whom the alleged illegal voters voted? …. Seems pretty shaky to me.”
Bingo! Katz even admitted that “if your assumptions are wrong, you get the wrong answer.” You get the wrong answer because the underlying assumptions are wrong (homogeneity assumption or exchangeability assumption).
I think Bridges will avoid making any broad-brush ruling on proportional analysis and simply say the Repubs’ evidence doesn’t satisfy the clear and convincing test.
Patrick; go visit a bank baby.
Worse for you, businesses also have processes designed to narrow the gap to ZERO. You know why? They lose their own money and sometimes (now w/ sarbanes oxley) might go to jail.
Tell me any other enterprise that willingly is inaccurate? Why isn’t Sims outraged?
The stink is pretty bad.
John@75,
By your own theory, all businesses should fail. Yet there are so many out there that are doing wonderfully and managing an inventory that makes King County look small. Boeing comes to mind. They are showing record profits. Do you know there are over a million parts in a boeing 747? How do they all get put together so that planes can fly? Magic?
Counting votes isn’t rocket science and no matter how many excuses you make it just doesn’t excuse the fraud that has occurred. The fraud is the dead votes, the felon votes, the illegal votes.
“Ever hear of the word shrinkage?”
Yeah, it is what happen’s to John’s brain when he gets on a blog. What happens when he goes into the water is another example but sink the thing shrinking wasn’t too large to start with, there isn’t much change.
The way the Republicans want to use the statistically questionable “proportional deduction analysis” by a cherry-picking a sample of illegal voters (mostly felons) in King County to overthrow a validated statewide election, I think they’d like to do ALL elections without elections. You can vote using statistics! Yeah! That’s it! No mess no fuss. No ballots. No lines. No machines. Just a few geeky statisticians manipulating data. Voilá! Perfect elections!
chew2 @ 74
“We disagree about what the GOP has to prove. Look, I’m a lawyer and it’s my opionion that the GOP only has to prove it for the precincts they have selected.”
We don’t really disagree about what the GOP has to prove. What we disagree about is the way the science can be used. I am a scientist, so I can only speak to the scientific issues. (Although that doesn’t stop me from expressing legal opinions ;-)
My question asked for a single precinct, and you said proportional reduction would work for that. You keep repeating that they have to show it for all the precincts. Why don’t you accept that we disagree about this.
You asked about 2 invalid voters in a precinct. I answered your question about what one could do for a precinct-level adjustment. If you want to do a statewide statistical adjustment, that same assumption must hold statewide. The votes to be removed must be a random sample (or complete census) of all invalid votes statewide. If not, the science part is incorrect whether it makes legal sense or not.
Here is another way to view it: the election total (129 in favor of Gregoire) is a statewide number. In order to change that number using statistical estimation technique, you must use assumptions that are valid at a statewide level. The precinct level only works at the precinct level, not at the statewide level.
“My question is can we demonstrate that Katz’s assumption of homogeneity is incorrect, and if so what does that do to his estimates and his estimates of error.”
Adolph and Handcock, in their reports, have shown that the homogeneity assumption is violated in the analyses done by Katz and Gill. A demonstration of this was Adolph’s simple example that ended up estimating that about 50% of felon voters were female. This establishes that the homogeneity assumption does not hold with respect to sex. It strongly suggests that the homogeneity assumption doesn’t hold with respect to voting (because we know that sex affects voting preference).
“My question is if you do account for that “uncertainty reflecting non-exchangeability” what can you say about Katz’s estimates and specifically about his error bounds. Can you quantify how much this increases the error. Or is all you can say is Katz’s study is flawed….or what?”
First, error bounds: without additional information (sex, race, etc.) the error bounds include the possibility that 0% to 100% voted for Gregoire. This is not rhetoric! So, rather than saying Katz’s method is flawed, I would say that it yields estimates with too much uncertainty to be useful.
With additional information like the (1) demographic characteristics (e.g. sex and race) of the felon voters, along with (2) some information on demographic characteristics of voters within precincts, along with (3) relevant exit poll data showing how the characteristics affect voter prefences, one can do a more sophisticated analysis. Handcock’s report does exactly that with sex (this is a different analysis than Adolph’s). Handcock adjusted estimates of illegal votes for Gregoire for the sex composition of the felon sample and the precincts, using exit poll data from the election.
You dems got a great scam going. Build a system designed to leak, let it leak, then challenge us to prove that secret ballots actually went our way. Orwellian.
Kind of like King County Elections adding the words “and record keeping” (ministry of truth anyone?)
DJ @ 86
I’m going to ignore your rehash legal analysis at the beginning. As I said we disagree.
But can you explain or reconcile these two statements you made:
1. “His (Katz’s) error estimates are fine if the only source of uncertainty in his analysis is binomial sampling error. In the binomial version of his statement (report 2?), for example, he properly has a variance in the final vote outcome that is 4 times the variance of a binomial, which accounts for the negative correlation between the scores as votes are removed from each candidate.”
****************************
2. “First, error bounds: without additional information (sex, race, etc.) the error bounds include the possibility that 0% to 100% voted for Gregoire. This is not rhetoric! So, rather than saying Katz’s method is flawed, I would say that it yields estimates with too much uncertainty to be useful.”
Katz claims in his report that the GOP wins, given his small variance estimates. (Granted he doesn’t take into account the Dem’s illegal votes.) Is your second statement a refutation of this claim by Katz. How would you use that second statement in arguing to Judge Bridges?
chew2 @ 88
Yes. I said in number 1 “if the only source of uncertainty in his analysis is binomial sampling error. . . .” In that case, he did quantify that component of the total uncertainty properly. I may have misunderstood your question when I gave that answer (if so, my apologies).
But, binomial sampling error is not the only source of uncertainty. In number 2, I was adding the uncertainty introduced by relaxing the homogeneity assumption.
“Katz claims in his report that the GOP wins, given his small variance estimates. … Is your second statement a refutation of this claim by Katz. How would you use that second statement in arguing to Judge Bridges?”
Yes it is a refutation. I would point out to the court that Katz’s variance is the binomial sampling variance only. If there is good reason to suspect that the homogeneity (or exchangeability) assumption is invalid, then the uncertainty of the estimates has been “underestimated.”
Both Handcock and Adolph make this point in their reports.
Goldy–
You were might “wiggly” on Carlson’s Show today. Several times when the discussion was about hard evidence, you changed the subject to relief and vice-versa. And you had no good answer for what the relief ought to be. Sharkansky sounded much more believable than you. If you continue down this path of ILLEGAL VOTES HAPPEN..RELIEF IS ASK FELONS & DEAD PEOPLE HOW THEY VOTED….I’m afraid Goldy you are destined to end up in HELL!!!
pbj, harry and joseph the incompetent troll(s),
You’re a riot, dude. It was your lame theory about the wonders of the private sector we were discussing.
How many dead people are we discussing? A couple dozen? Is KC the only county that had dead people vote? Is KC the only county that had felons vote? Hmm. maybe so except for those 8 in Chelan and some in Spokane and some in Yakima and some in Ferry, etc, etc. Let’s frog march and hog tie those elections officials in their PRO-ROSSI counties too!
Keep spinning your paranoid fantasies about King County. The more you spin, the more I laugh – at what passes for your lame brain!
pbj @ 84
time for another episode of RCW 101 class. Fraud committed by voters has no relief in Washington courts. The only remedies are for official fraud, for which evidently there is none, or the GOP would have produced the evidence at trial.
Go for it John, because the bottom line is the anomalies in Kiev County (soon to be quite a bit smaller) add up to incompetance on a grand scale – you smaller county Kievers (read, Seattle libs) will get exactly the government you deserve.
Totally wrong @ 83
Wow, dude you’re really into the accuracy of banks. All elections outfits should run just like those banks. If not, fraud, jail time for those awful people running elections.
So how about the Elections bank of Chelan who let 8 felons vote?
Or the Elections bank of Yakima County that let x number of felons vote?
Or how about the Elections bank of Spokane County that let x number of felons vote?
Or Ferry County?
etc, etc, etc…
Fraud, Fraud, Fraud, jail, jail, jail goes the unthinking, winger echo-chamber script…
John,
I’ve never engaged on the felons, except as a bit of Chinese water torture on the boobs at KC elections. Real issue is the excess votes in “special” precincts, and the take home ballots that KC fostered.
Felons will occasionally vote, same w/ a dead spouse or two.
Lets say we GOP lose, will you guys ever then support cleaning up our elections (strip away the partisan fight part of this). You really want smiling Dean logan to keep lying about his great recordkeeping, Ron to lie about is 99.8% accuracy.
And yeah, we should expect accurate AND honest election. We likely all have relatives who fought for a variety of American causes/traditions, one being the right to vote (which implies an honest vote that counts (not mine that your guys wiped out by feeding an extra ballot into the Accuvote)
righton: If the “take home ballots” are indeed the “real issue”, when will we be hearing about them in court?
I don’t think we have that chance; (not a lawyer, not that close to it). I think just showing lots of errors and gaps and holes and all that is not enough.
Question i keep asking is how big does the bucket of leaks and monkey business have to be before some authority says, ok, boys that is wrong, stop doing that.
You all are so hung up on Gregoire that you’ve never said this thing stinks. Get sims and logan to admit theft and i’ll give you gregoire, really (she’s darn good for the gop for the next 3 odd years)
righton: I think the judge is allowing Rossi remarkable leniency in allowing his lawyers to put whatever case they want before the court. If they don’t present something, I can only assume that it’s because there isn’t much to it. As I noted, they don’t seem to be saying anything about your “take home ballots”.
My bad, Sam Reed, not Ralph :)
But many of you on the right are missing the point.
THERE IS NO ANONYMOUS ELECTION THAT HAS EVER BEEN HELD WITH AS MANY VOTERS AS THIS ONE HAD THAT WAS 100% ACCURATE.
This election was a TIE. They might as well have flipped a coin.
The election statute was followed and Gregoire came out on top. It could have just as easily been Rossi that came out on top and we’d have the Democrats talking about how horrible the King County Elections commision was.
I’ve LIVED in cities where the election contests were routinely messed up. Washington State’s numbers are a paragon of virtue compared to some of the places I’ve lived (I remember hearing a story told by one of my dad’s co-workers that the local Republican ward boss called her up one day after a presidential election to ask why she had voted for the wrong party).
THAT’S an example of what you get when you have a broken election system.
It’s technically NOT POSSIBLE to get 100% accuracy when you’re holding an anonymous election for 2.5 Million voters.
You can probably do it if you remove the anonymity in the voting process. But as long as there is no relationship between the voter and their vote (which is a critical part of our democracy), you’re not going to have perfection.
It just doesn’t happen. And it can’t happen.
Totally wrong @ 95, 97
You have as much proof of foul play from Logan and Sims as I have proof of monkey business from the folks in the pro-Rossi counties with the various errors they made, i.e. none.
Sam Reed is putting in his database to weed out the felons and the dead people. Satisfied? Didn’t think so.
The provisional ballots are probably going to be color coded. In the last special election, a sticker was placed on them so the accuvote machine would reject them. Satisfied? Didn’t think so.
Just about every step in the processing of votes is open to scrutiny from the public and all the political parties. Satisfied? Didn’t think so.
The only thing wingers will be satisfied with is a re-architecting of voting methods and procedures that is designed to put more wingers into office period.
If you take us back to the ballots that were counted at the first go around, and count the crap out of that pool of ballots i’m with you.
But once you start injecting ballots at the same time your are counting, and you also really don’t know who has the ballots, then NO WAY is this the most accurate large election
Nice try, I almost bit, but its the “count and inject” thing i cannot get past.
John,
Holding up Sam Reed as an election guru or paragon of counting is a joke. He’s clueless.
Yeah, lotta reform. ZIP. and now more mail in voting. whoa, think of all the homeless guys stuffing ballots next go around.
Gotta get that doctor in Atlanta his ballot early
PS I have testimony in court of a conspiracy to fudge documents. And evidence in court of non-sensical counting of ballots.
by the way, anyone wanna tell me why they let workers take blank ballots home?
Totally wrong @ 101
Many counties beside King “found” votes during the recounts. All of these votes were scrutinized by all the parties. There is no proof of ballot stuffing. The case the R’s are making in court so far has produced no smoking gun.
If you have proof, please link to it.
Just go read the Michelle Way stuff. Try the transcript at SP, not goldy’s softpeddle. She and her boss (btw she’s a supervisor so already a boss), signed a report that was false. Lib spin is she didn’t know it was bad at the time, but she actually did very soon after. AND ignorance per statute doesn’t excuse her conduct.
Many counties beside King “found” votes during the recounts. All of these votes were scrutinized by all the parties. There is no proof of ballot stuffing. The case the R’s are making in court so far has produced no smoking gun.
Yes, but their case is not over yet. There still might possibly be weapons of mass destruction that justified this election contest.
Righton @87. Nice post. Could not have said it better.
I don’t know if the lib posters are being deliberately obstinant for political gain or if they actually believe what they are posting.
I believe Burman ripped apart Katz’s nervous testimony by using authors Katz cited as his background, who called his method bogus. I believe the statistical principle Katz claimed to rely on did not appear anywhere in the book Katz cited. I believe Chris Adolph laid out two quite obvious problems with Katz’s assumptions, and proved the second one straight out with the comparison of precinct females to felon females. And I believe Bridges himself all but told Braden flat out that he was coming with nothing on ‘clear and convincing.’
And I believe Rossi has less than 24 hours to keep Bridges from being sorely tempted to drop this loser altogether (although he’ll likely make the Democrats give their side as Supreme Appeal insurance).
I believe Phase II of the War on Gregoire is about to begin for the Washington Right.
We dare you all to put Deano up on the stand. Barney Fife in a suit.
Righton: Why don’t you examine the evidence for vote fraud in Snohomish County as it relates to the touch screen voting machines?
I’m not familiar with any proof of that baseless allegation. Do you have testimony, evidence, grand jury, election supervisor affadavits of fudging reports?
is this Harry Reid ; you should be blocking votes in the Senate
So should we all just read David Postman? Is the the Goldy blog or the “David Postman” blog? C’mon Goldy. You can do better than that can’t you? Entertain us.
Joe@92,
Perhaps but at the very least you should have the guts to admit it was anything but a “model election”. Any fair minded person would at least conclude that much.
youe really need to shape up ur language!!!