Live by the cherry-pick, die by the cherry-pick.
In court documents filed before Friday’s deadline, Democrats claim they have identified a total of 743 felons who voted illegally, mostly from pro-Rossi precincts. This is not only more than enough to offset the 946 felon votes claimed by Dino Rossi’s attorneys, but using the GOP’s very own statistical methodology, may even expand Governor Christine Gregoire’s 129-vote margin.
Washington State Democratic Party Chair Paul Berendt also claims that this is evidence that Rossi deliberately concealed evidence of illegal votes cast statewide:
“The discovery of these additional illegal votes reinforces the Democratic Party’s long-held belief that Dino Rossi has concealed illegal votes that benefited him in November.
This new information explains why Dino Rossi argued that evidence of the illegal votes that benefited him should not be considered by the court, because if you look where Dino Rossi wants the court to ignore, he loses even under his own theory.”
The Democrats issued a press release with a map showing 22 counties where they claim Rossi concealed names of felon voters who would have benefited him under his own proportional reduction theory. The Democrats found additional felons in other counties, in precincts that also voted pro-Rossi.
| Counties Where Rossi Concealed Illegal Votes |
 |
According to the Seattle Times, Democrats have also added two University of Washington professors to their legal team, who plan to argue that felons are more likely to have voted for Rossi than Gregoire.
“We know for a fact that nonunion, blue-collar, Caucasian men vote very disproportionately Republican, and when you look at the felon population in the state of Washington, they are overwhelmingly nonunion, blue-collar, male Caucasians,” said state Democratic Party Chairman Paul Berendt.
“This could be under any other circumstance the prime demographic Republicans target in their” get-out-the-vote effort, he said.
This is an issue we’ve argued forwards and backwards in the comment threads, and I’m guessing we’ll argue it some more. But the point is, if you’re going to settle this election by statistical analysis, you’ve got to make sure you’re analyzing the right statistics. For example, 95 percent of felons are male, and thus it would be misleading to simply infer that precinct is the only significant predictor of how they may have cast their ballot. Republicans would like you to believe that the real winner of this election can be determined by a simple math equation… but nothing is further from the truth.
Anyway, here’s the final county breakdown of felon votes alleged by Democrats:
| Adams | 1 |
| Benton | 53 |
| Chelan | 21 |
| Clallam | 12 |
| Clark | 57 |
| Columbia | 1 |
| Cowlitz | 10 |
| Douglas | 7 |
| Ferry | 6 |
| Franklin | 8 |
| Garfield | 1 |
| Grant | 45 |
| Grays Harbor | 1 |
| Island | 5 |
| Jefferson | 1 |
| King | 58 |
| Kitsap | 31 |
| Kittitas | 8 |
| Klickitat | 3 |
| Lewis | 22 |
| Lincoln | 5 |
| Mason | 1 |
| Okanogan | 11 |
| Pacific | 1 |
| Pierce | 86 |
| Skagit | 17 |
| Skamania | 1 |
| Snohomish | 52 |
| Spokane | 86 |
| Stevens | 1 |
| Thurston | 29 |
| Walla Walla | 11 |
| Whatcom | 8 |
| Whitman | 7 |
| Yakima | 76 |
Of course, Republicans are already questioning the accuracy of the Democratic list, and I’m sure some of these names won’t pan out… but then, neither will some on the Republican list. I think it is telling that with all the GOP talk of shoddy work on the part of the Democrats, Vance could come up with very few actual examples in the Times.
State Republican Party Chairman Chris Vance said that 31 of the people included on an early Democratic list were among the felons listed by Republicans. He also said he doubts the accuracy of the Democratic claim that residents at Western State Hospital voted illegally because they had been found mentally incompetent. Vance said Republicans investigated those votes earlier and were unable to determine whether they were illegal.
“It feels to me like the Democrats are desperately trying to build a last line of defense,” Vance said.
The fact is, Rossi would need at least a 700 felon advantage (if you can call it that) to have a hope of prevailing on his most favorable proportional methodology… so the Democrats have plenty of cushion. And if Vance wants to talk about desperation, he should take a look at the comment thread over on (u)SP where they’re finally analyzing the transcript of Judge Bridge’s recent rulings, instead of just reading the AP headlines. Some of the most ardent and articulate defenders of Rossi’s lawsuit, are pretty much conceding it’s over.
I haven’t had a chance yet to more than glance at the transcript myself, but the previews sound very encouraging. More later.
UPDATE:
A clarification. The Democrats press release describes its list as “illegal” votes, not “felon” votes. Apparently, the total number includes all of their alleged irregularities, including ballots that should have been counted, but weren’t. The list identifies the voters by precinct, and the court should make it available to the public by Monday or Tuesday. Looking at the numbers, it would seem that like the R’s, the D’s are not including mishandled provisional ballots in their list.
UDATE, UPDATE:
Okay, I’ve been corrected. The list submitted by Dems identifies 743 probable felon votes in 35 counties. It also identifies 1898 provisional ballots counted before verification in 14 counties, not including King. And there are an additional 780 absentee and provisional ballots from King County that are identified as not being counted due to error by election officials.
My bad.
Goldy–
So Berendt is NOW offended by cherry-picking???
What a turnaround for Bawling-Pauly!
What about the provisional ballot affadavits where the Dems ADMIT going to voters, asking how they voted and then if they said ROSSI….Berendts COMRADES said “Have a nice day!”. If they said GREGOIRE, one of the LEFTISTS rushed down to get a new signature card and a signed affadavit (and a number of these affadavits are undoubtedly forged!!…remember the $10 check validation procedure??). And Berendt has the gall to get in front of that camera sniffling and wailing that he spent his whole life “wanting to make sure every vote was counted”. BULLSHIT!!! If that wasn’t cherry-picking…what is??
Hey Goldy—the cards are now on the table for this one issue.
Evidence will be evaluated…these lists will be pared down. The Judge will listen to arguments from the “experts” on the allocation methodology. The formula the Judge accepts will be applied to the final sanitized list of felons. It ain’t that complicated.
Oh and Goldy—think long and hard about the word APPEARS.
Perhaps there are numerous statistical formula’s in the acceptable range that the Dems and R’s present for proportioning these felons. In other words, there is no one PERFECT formula, is there? No. The R’s must show that it APPEARS these felons cost ROSSI the election. APPEARS Goldy.
But it does appear the Dems may have made a last minute stab and possibly removed a couple of the names that were obviously legal votes. How will this contest wind up??
It depends on the final sanitized list of felons accepted by the Judge AND the proportional analysis formula he feels meets the standard of APPEARS.
And it also depends upon how the Judge rules on the other issues still on the table.
No one gets to excited about a gasbagging blowhard like Paul Berendt. He is the kind of guy that screams “Support our Public Schools!!” and then takes away over $8000 from his own local school by sending his 2 kids to an expensive private school in Olympia.
Ah, how darling, felons vote everywhere! Not just King County.
All those corrupt elections boards everywhere committing distributed Republican vote fraud. Dean Logan is in good company – in the Red Counties.
It just warms the heart doesn’t it?
Gotta love that canvas of the provisional ballots by the Dems; what possible reason do you allow that, unless the goal is to jam some more Fraudoire votes into the box
righton@3
Right on!
I wonder what DJ’s latest county-level proportional analysis of the average expectations and statistical likelihoods is now, with the Dems latest felon list being added into the mix.
The felon vote argument may (or may not) win the day for the GOP, but it does lack quite a bit in moral force. There is not evidence that anyone made any organized effort to get 1600 or 1700 ineligible felons to vote in order to favor one side or another. These felons are all real people, and they registered to vote openly. If anyone had cared enough, they could have been challenged before the election. If illegal felons does the trick, then it is almost like winning on a technicality.
I am much more offended (and in an extreme way) about totally fraudulent votes that could not have been prevented by due diligence before the election. Such as people voting twice, people impersonating the deceased, people who weren’t regustered to vote, and phantom extra ballots that were not cast by real people.
For example, King County’s voter crediting records still show around 2,000 more ballots than voters in November 2004. Close to 800 of these can be explained by irregular provisional ballots. Of these close to 800 irregular provisional ballots, something like 170 or 180 of them were not cast by registered voters (or represented people voting twice).
This leaves another 1,200 or so excess King County ballots that cannot be identified to any particular person (at least so far). Granted, this could be the result of voter crediting errors. It would be possible for the GOP to have people go over the names of all 900,000 or so voters whose ballots were accepted — 305,000 or so poll book signatures, 565,000 or so accepted absentee envelopes, 28,000 or so accepted provisional ballot envelopes, and 800 or so irregular provisional voters — make copies of these records, and then add the numbers up appropriately. It might cost $100,000 to $200,000, but that is chump change in comparison to the overall cost of this election contest.
My guess is that King County would still have several hundred extra unexplainable ballots, even when the “best evidence” of voting is considered, and all the actual people voting are counted and added up. I think if you have ballots that have really been “stuffed”, and due to official error or fraud (as opposed to the act of an individual ineligible voter), then it should be sufficient simply to show that 130 of these completely mysterious ballots exist in order to overturn the election.
Trouble is that there seems to be no evidence that the GOP is still trying to prove that voterless ballots were cast in King County — in spite of all the hoopla over this and the obviously offensive nature of simply having extra ballots stuffed into the mixture.
“””Trouble is that there seems to be no evidence that the GOP is still trying to prove that voterless ballots were cast in King County – in spite of all the hoopla over this and the obviously offensive nature of simply having extra ballots stuffed into the mixture”””
Well, in the spirit of the republican party; maybe the reason they aren’t pursuing this, is the same reason they ignored eastern wa. felon votes.
Exposure.
It was a big surprise to see rossi tie gregoire.
Maybe, there’s more to it?
AllHat–
The Republicans did not “ignore” E. Washington
This is an Election Contest.
It’s like asking a defense attorney to prove his client is guilty.
The Dems have made themselves a Party to this contest.
They have now given it their best shot.
Let’s see what happens.
That said, I agree with Richard Pope about the more offensive offense is the voterless vote issue in KingCo and the bogus Canvassing Board certification. These issues are not dead but….
There is no doubt in mind mind that given the KingCo reconciliation efforts, lack of internal controls and procedural meltdowns that there was ballotbox stuffing.
Also–
Don’t you think the Dems may have come across some felons still not on the list that turn out to benefit Rossi???
Just like the unseemly spectacle the Dems participated in interviewing questionable provisional ballot voters about who they voted for (post election) and then only pursued affadavits for those who voted for Gregoire.
AllHat===selective ethics dude.
It just proves what Rossi said “this election was a mess” was right on.
6 months after the election ‘The berendt’ looks like a lying cry baby after saying ‘every vote should count.’ Gregoire looks like a lying rat by saying ‘this was a model election.’ And Ron Sims looks like a good home trained liar for saying ‘any bank would envy.’ Will your party ever stop embarassing themselves?
‘The Berendts’ latest spin that “nonunion, blue-collar, male Caucasians” vote Republican is BS. It just so happens that there are more “nonunion, blue-collar, male Caucasians” than union, white-collar males of all races.
Is ‘the berendt’ related to James Carvelle?
Cynical, Pope, & Chardonnay,
Can you three honestly say that you would have been so concerned/outraged/whatever about illegal votes if this election hadn’t been so close?
If the margin of victory were not less than the margin of error, you wouldn’t have given these issues a second thought. Especially so if your guy had won.
Forgive me if I find the moralizing a little disingenuous.
And Cynical,
You’re all for letting the chips fall where they may. Why didn’t that same logic apply to the recounts. In the case of the recount, at least there was a well-defined process to follow. In the present court case, they have to make up the rules as they go along. Seems much less “conservative” to let the courts decide, rather than following the laws created by the legislature in the first place.
Looking for signatures to validate ballots, is one thing.
Looking to supress known illegal votes, is another.
The repubs. thought they would claim the “BIG MESS” theory, by showing illegal votes, then when it turns out, there were many fewer errors in this election, than other similar races, they chose to use their “too many felons voted for gregoire” theory.
Face it, they fucked up. every election, has similar errors.
Im as much or more concerned as to whether or not, there were other reasons for repubs. to distract everyone with these issues.
Maybe we’re overlooking something.
AllHat–
Berendt and the Dems suppressed known Rossi provisional ballots in the affadavit scam.
It’s a contest AllHat.
The R’s had no obligation to canvass the entire state.
The D’s elected to become a party to this lawsuit.
Now they have looked for offsetting ballots.
Where is the harm???
Your comment “looking to supress known illegal votes”….
can you prove that AllHat????
I didn’t think so.
Simply looking only at Gregoire precindts was a reasonable strategy validated by the Judge who allowed Dems to “cherry-pick” provisional affadavits weeks after the election.
GIMMEE A BREAK!!
You AllHat “Spin” won’t fly with the Dems prior provisional ballot affadavit scam. Bad try===No sale!!
Rtoes–
Ummmmmm, Dumbass sir…
Do you have an alternative to letting the chips fall where they may??
The law allowed the 3 prior counts.
The law allows an election contest.
We are at the crossroads of the election contest.
The evidence is on the table.
It will be validated and the Judge will rule.
Let the chips fall where they may!
Cynical,
Wow. Name calling. That’s mature.
The recount was automatic.
The court challange is not. R’s chose to take it to court. They could just have easily chosen not to.
My point was the following: conservatives crow about the arrogance of the judiciary, except in the situation when it might protect them.
Ummm. Hypocracy?
rtoes–
The law allows for an election contest.
Gore did it in Florida didn’t he??
You are saying that 3 million voters should simply trust what Dean Logan, an APPOINTED Democratic Party political hack Election Director and Ron Sims “little fella” came up with?? With no legal review???
The other County Auditors are ALL Elected Officials.
Logan is APPOINTED.
Huge difference, don’t you think?
I dont have to prove it. It will be obvious in court, especially with vance whining, he knew all about the alleged illegal votes the dems. now claim.
Again, to deliberately not tally illegal votes, because they favor the other team, is deceptive at best, and wont sit well with bridges.
toes is also correct about rossi wanting to fast track, while his lead was slipping.
There was plenty of effort by repubs. to stop the counting, and move on.
How many diebold machines were used in wa.?
Were they predominantly in strong repub. precincts?
AllHat-
Calm down little fella…calm down.
Now you are bringing in a Diebold conspiracy theory??
Sounds desperate.
Go ahead and continue to rant & rave about the R’s not scouring the state for every felon and how horrible that is.
If you didn’t notice, for some strange reason the Dems did it.
Hmmmmm?
Kind of shoots your argument in the butt, doesn’t it?
So what you are saying is.
“Cherry-picking provisional ballot affadavits by calling voters after the election is over, asking them how they voted and going out and getting the Gregoire ones (while hanging up on the Rossi ones) is OK”???
While each Party making the effort to build a case for felons in their opponets precindts is not ok???
Great logic lads!!
The diebold theory, has at least the same credibility, as did the “call all dems. ballot stuffers” theory.
What is the crime you suggest dems. are guilty of, by hanging up on rossi voters?
The one I suggest, when repubs. denied tally of illegal rossi votes, is withholding evidence, and conspiracy.
Richard @ 5
I’ve done the Monte Carlo analysis at the county level using the new data. For the three analyses given below, I simulated 100,000 trials of “proportional reductions”. For each trial, invalid ballots are randomly removed from the county totals according to the proportion of Democratic, Republican and Other voters. I assign ties randomly between the two major candidates (actually, I should give them to Gregoire because that is the most likely outcome of a tie in the election; but, my procedure is conservative for Rossi).
Analysis 1
Democrats win: 95008 simulated trials, Republicans win: 4992 trials
Democratic wins: 95.0%, Republican wins: 5.0%
Mean final vote margin: 71.0 for Gregoire (SD: 42.3)
This, of course, assumes that all 1954 illegal votes by all sides are accepted by the courts. Clearly, the Republicans have not proven that Gregoire owes here victory to errors under this scenario.
Analysis 2
Suppose one half of the illegal votes on the Democrats list are invalid, but all of the Republican list stands. Who wins then?
Democrats win: 76660, Republicans win: 23340
Democratic wins: 76.7%, Republican wins: 23.3%
Mean final vote margin: 23.7 for Gregoire (SD: 37.3)
So, even if only one half of the ex-felons on the Democrats list are accepted in court, there is no chance for Rossi to prove that Gregoire owes her win to illegal votes.
Analysis 3
Suppose only 10% of the Democrat’s list and 100% of the Republican list is accepted.
Democrats win: 33057, Republicans win: 66943
Democratic wins: 33.1%, Republican wins 66.9%
Mean: 13.1 for Rossi (SD: 34.2)
So, even if Democrats have made massive errors in their list (90%) and Republicans have made no errors in their list, the results only give about 70% certainty. The big question becomes, is 70% certainty enough to overturn a certified election? The Judge would have a 30% chance of erroneously overturning the election. And, at this level of certainty, have the Republicans really proved that Gregoire owes her victory to the errors? Probably not. This level of significance is never acceptable in science, where the stakes are much lower. I have argued elsewhere that a level closer to 99% is more appropriate for overturning a certified election.
dj,
How different in results is your monte carlo simulation from the Katz method using county data. My recollection is that using his county data your result was less favorable than his to the republicans, although the basic theory behind both is the same.
Was this difference due solely to his use of the normal approximation to the binonmial, or is there something else at work.
Using the precinct data over the county data, Katz gets a much more favorable result for Rossi. Is there any way for you to estimate what the result would be using your monte carlo method.
I can’t wait to see how bad the margin of error is on their preliminary findings. I would venture to say it will be as bad if not worse that Republican findings.
Bottom line is that both sides are getting over zealous. Goldy screams that Republicans don’t have a case, Sharkansky screams that Dems are losing the case. partisan political blogs don’t have all the answers, although I will say Sharkansky provides more fact to back his claims even if you always make fun of him for it (pdf files and spreadsheets).
Mr. Cynical and AllHatAndNoHorse are having an interesting argument about “cherry picking” at two different points in this process.
The first point was when KC released a list of voters whose provisional ballots needed an affidavit. As I recall, Democrat-leaning workers went door-to-door trying to seek affidavits. Some of the workers would first ask whether the person was a Democrat or a Republican. If they said �Republican,� they would politely move on.
Personally, I find this distasteful. But is it illegal? Clearly not! The list of names were provided to both parties. If my memory serves me correctly, the Republicans decided to use a low-labor approach of calling people and telling them to go sign an affidavit at a nearby courthouse. The Democrats used a much more aggressive strategy. They had a right to do so, and in this case, the Republicans simply fucked up by not doing the same thing.
If KC had pulled the same thing, however, it would clearly be election fraud. The county could not legally cherry pick. But, KC did not undertake re-contacting the voters, and had no control over the way in which any third party went about contacting voting.
The second point has to do with identifying illegal votes across the state. Cynical writes:
The R�s had no obligation to canvass the entire state.
Indeed, the Republicans had no obligation to canvass the state. But, was this a good legal strategy? The Republican strategy hinges largely on the acceptance of a proportional analysis. This means getting statistical-types involved and using a certain amount of statistical reasoning. It means statisticians will be providing evidence. I can guarantee you that any statistician put on the stand will testify that a proportional analysis is only valid if all counties undergo the same degree of scrutiny.
Thus, the Republicans can cherry pick all they want, but their case is, statistically, a joke unless they can show uniform standards across counties and precincts in uncovering invalid votes. There is just no way around this.
The Democrats are now doing their own cherry picking. But, the Democrats have an entirely different role. They have the luxury of cherry picking because all they must show is that the Republicans did not apply uniform standards across counties and precincts. I believe the Democrats have no interest whatsoever in using their illegal votes in a counter-proportional analysis. Their strategy will be to show (if a proportional analysis is ruled a valid statistical methodology for the court) that the way invalid vote data were collected make the Republican proportional analysis statistically invalid.
The Republicans really fucked this one up irretrievably by not applying equal scrutiny. On the other hand, if they had done so, they most likely would have found that the illegal votes canceled each other out in a proportional analysis. They could have dropped the lawsuit and saved a bunch of money.
DJ@23 says,
“”The Republicans really fucked this one up irretrievably by not applying equal scrutiny. On the other hand, if they had done so, they most likely would have found that the illegal votes canceled each other out in a proportional analysis. They could have dropped the lawsuit and saved a bunch of money.
Comment by dj””
This is imo, exactly what happened.
They started out with the “all messed up” theory, then went to the “felons voted for dems.”, theory, Then realized, oops! they voted for rossi too!!
This led to the cherry picking, as they were already too far down this road to backtrack to another completely different avenue of accusations, and hoped, this would be enough to cloud the election results.
Its not, money was wasted.
Chew2 @ 21
How different in results is your monte carlo simulation from the Katz method using county data. My recollection is that using his county data your result was less favorable than his to the republicans, although the basic theory behind both is the same. Was this difference due solely to his use of the normal approximation to the binonmial, or is there something else at work.
Yes, the basic theory is the same. He is using a method that is supposed to approximate the results that I get using a Monte Carlo simulation. There are a number of weaknesses in Katz’s analysis versus the Monte Carlo simulation, however.
1. Katz’s analysis apportions bad votes only to Ds and Rs. This is implicit in using a binomial distribution. In fact, adding an “Other” category for third party, independent, and undervotes makes Katz’s analysis less favorable to Rossi. Given how close the results are using the original Republican list, Katz should have used a trinomial distribution. (He speculates that it is not important, but I showed that it really is).
2. Katz uses a normal approximation to the binomial distribution, relying on the fact that the areas under a binomial distribution can be approximated by a normal distribution under some circumstances. This causes a bit of error. The interesting thing is that the exact areas under the binomial (or trinomial) distribution are not all that difficult to compute. He could have done that instead.
3. Katz’s method implicitly assumes that there are no ties. He is using a continuous approximation that logically accepts a win by, for example, 0.001 votes. The simulation method works with real (simulated) ballots, so there is a “tie” category when the final vote tally is 0. I have been splitting ties between Rossi and Gregoire. But, if the election had been tied, a joint session of the legislature would almost certainly have elected Gregoire. Therefore, ties should go to Gregoire. (An alternative argument is that, since Rossi must prove he won, ties should favor Gregoire).
4. Katz’s method does not consider that once a ballot is removed, it stays out of the pool. Statistically, the Binomial distribution is “sampling with replacement”. Here is the difference: suppose we physically went through the trouble of randomizing all the ballots in each precinct and randomly removed a quantity equal to the invalid ballots. Under the assumptions of Katz’s analysis we would 1. remove a ballot. 2. note who the vote should be docked from. 3. throw the ballot back into the pool of ballots. 4. go back to step one if there are more invalid votes. I set up the MC simulation so that it does not do that. Instead, step 3 is skipped (the selected ballot is set aside). This make a difference at the first decimal point, but that would have been important with the original Republican list.
Now, with the list of invalid votes found by the Democrats, the subtle differences between my method and Katz’s method are less important. Katz’s method rules out the possibility that Gregoire owes here win to the errors. And, I can state with great confidence that this is true, even if done at the precinct level. The Republicans will have to knock out 80% to 90% of the invalid votes on the Democrat’s list to have even a prayer.
Using the precinct data over the county data, Katz gets a much more favorable result for Rossi.
Using the precinct analysis has the effect of reducing the variance a bit. This favors whoever has the higher average number of votes in the final tally. So, if the average election outcome is in favor of Rossi, he is favored. If the average election outcome is in favor of Gregoire, she gets an advantage from a precinct level analysis.
Is there any way for you to estimate what the result would be using your monte carlo method.
The best thing would be to get a copy of the precinct-level data. Any help with that would be appreciated. I may be able to estimate a fudge-factor in the mean time (if I can find the time).
Cynical @1,
Again with the “APPEARS” defense. Well, if that’s the best your BIAW buddies have, you and Rossi and the GOP could have saved yourselves a few million dollars and just bought yourself a copy of Black’s Law Dictionary, which defines “appears” as: “To be in evidence; to be proved.”
Looking at the various parties’ filings, and reading through the transcript of Judge Bridges’ rulings, I’d say it appears Rossi is going to lose his contest.
Goldy–
You are the wordmeister!
dj–
Try applying your simulation on a “precindt” basis.
Oh, that’s right, you can’t because you don’t know what precindt they voted in.
Allhat–
OWWWWWWWEEEEEEE! Now it’s withholding evidence and conspiracy the Republicans are guilty of. That sounds pretty bad, doesn’t it?
dj,
Thanks alot for the explanation.
Two more points/questions.
1. My understanding is that the precinct level data may be more favorable for Rossi because the Gregoire plurality in those precincts was higher than the county plurality as a whole.
2. As I understand it using ecological inference statistical techniques it is possible in principal to estimate proportion of males who voted for rossi or gregoire, if you know the sex breakdown of voters in a precinct and the vote results for the candidates. This estimation is apparently done for race in voting rights cases to show whether blacks tend to vote for black candidates, and whites tend to vote for white candidates.
Since we know that most of the felons were male and in fact can identify each and every one of them, would it be possible to estimate the male propensity to vote for Rossi, and factor this in to an estimate of how the felons voted. I’m not sure whether the sex of the voters is known for each precinct however. But somehow the researchers are able to get race and sex data for voting rights cases.
Do you have any comments about ecological inference?
Richard Pope @5
I agree with you on much of your analysis and sentiment. However, I again would like to point out that the voter-credit discrepancy is not evidence of stuffed ballots, and is most likely the result of data entry errors during the voter credit process.
I’ve touched on this before, but my main problem with the theory that election workers stuffed hundreds of ballots, is that this would require both a mechanism and a motive for doing so, neither of which have ever been suggested.
By motive, I don’t just mean that these workers might prefer to see Gregoire win, but that they were willing to sacrifice their job and their liberty towards this effort, while having a reasonable expectation of impacting the final results. Considering the extreme scrutiny of the counting process, any such effort would have had to occur during the initial count, and likely on election day… but certainly before the final results were known. As such, why would anybody risk all simply to stuff a couple hundred votes for Gregoire when she was expected to win by nearly 100,000? Even after election day, nobody knew it was this close. Indeed, if you look at Stefan’s projections (and mine were in total agreement), we all thought Rossi would win by over 3000 votes. So again, why bother stuffing a couple hundred ballots that wouldn’t change the outcome? (And in the context of the first two counts… didn’t.)
As to mechanism, nobody has ever addressed this. How exactly does one stuff a couple hundred ballots? How does one obtain the blank ballots, mark them, sneak them into the counting room, and randomly distribute them amongst King Counties 2616 precincts? Because if you look at the voter-credit discrepancy, it does not occur all in one place… it’s mostly scattered in onesies and twosies across the county.
What you are suggesting is that one or more election workers or Democratic officials brilliantly conspired to stuff ballots in a virtually undetectable way… yet were too stupid to do it in a quantity that had a reasonable hope of swinging the outcome of this election… which was decided by an unexpected and anomalously narrow margin of only 129 votes.
I agree with you that if there was evidence of several hundred voterless ballots of unknown provenance, that might be enough to toss out this election, as it would be evidence of fraud… something that was proven in the Foulkes case, but lacking from this one. But as Judge Bridges’ has ruled, the voter credit discrepancy is not admissible as evidence of illegal votes… at the most, it represents a lack of evidence that there weren’t illegal votes. So without some evidence that ballot stuffing actually occurred and how it was done, I just can’t believe that ballot stuffing was likely. It doesn’t make sense.
As to whether the GOP has tried to find evidence of such activities, it was my understanding that they did examine a number of poll books and compared them to the voter credit records. I assume this didn’t pan out.
DJ @20,
Thanks for quickly running new simulations. We should have precinct-level data available sometime next week.
That said, you’ll have to program another twist into your analysis, as the Dems’ list includes legal ballots that should have been counted, but weren’t.
DJ @23,
Great analysis. As I’ve pointed out before about Stefan’s work, your mathematical formulas are worthless if your data sucks. Garbage in, garbage out.
I have talked to Dem attorneys, and you are absolutely right, their goal is to discredit the R’s methodology, not counter it with offsetting votes. But they’re convinced they win either way.
Oh… and we should have the precinct data in a few days.
AllHat @24,
I think the flaw in the Republicans strategy from day one is that they believed their own PR: that Democrats are crooks who tried to steal this election. There were many in the Rossi camp who believed that if they looked hard enough, they would uncover evidence of fraud… thus they were never focused on the proper objective… winning a contest on irregularities.
Goldy @ 29
I will preface my comment with the legal reality of what it would require to prove ballot stuffing (in the absence of direct eyewitness testimony). The GOP would have to go over all the original source documents (poll book signatures, ballot envelope signatures, and the documentation regarding irregularly cast provisional ballots) and RELIABLY AND BEYOND A REASONABLE DOUBT (or at least to the highest civil standard of clear and convincing) enter this information into a database.
This would probably take between $100,000 and $250,000, but would produce reliable and convincing evidence. (Of what, we wouldn’t know unless it was done. It could end up be convincing evidence that no ballot stuffing occurred.) It doesn’t look like the GOP has done this or is doing this, so this probably will not be part of the election contest case at trial.
That being said, there are plenty of things which cause me to have a lot of suspicion. The continued failure to resolve the voter crediting errors is a major problem. Not sufficient evidence by itself, but enough to cause a lot of suspicion and investigate further.
More alarming is the unexplained increase in ballots counted in each of the three counts.
In the original count of 11/17/2004, 898,238 ballots were counted.
In the machine recount of 11/24/2004, 898,574 ballots were counted. NO BALLOTS WERE OFFICIALLY ADDED TO THE MIX. Where did the EXTRA 336 BALLOTS come from then?
In the manual recount of 12/23/2004, 566 ballots were officially added to the mix (the “Larry Phillips” misplaced ballots, and maybe two or so absentees locked inside machines). That should have boosted total ballots counted to 899,140. However, the manual recount actually counted 899,199 ballots. Where did the EXTRA 59 BALLOTS come from then?
Basically, after the initial (and normally final) count on 11/17/2004, we have a total of 395 EXTRA BALLOTS counted in King County that simply cannot be explained — more than three times Gregoire’s purported victory margin.
Again, it is a pity that the GOP apparently isn’t spending the money necessary to compile a list of actual voters from the original source documents. If this was done, and there was a discrepancy of more than 129 ballots, then the election would have to be nullified based on fraud.
Goldy–
How are ballots that should have been counted but weren’t….illegal ballots???
Could it be the Dems now want to count those extra ballots found in envelopes in KingCo???
If so, what is the rationale for that???
Goldy–since you are reporting it, maybe you can do a little better job of explaining the Dems mighty 743!!
dj – Did your simulations take into account the fall-off (by county) of voters who left the Gov race blank?
There’s no way in evidence to literally satisfy Judge Bridges’ that dq’d felon votes don’t enter the case unless they voted for Governor, so I assume this envisions proportional treatment of felon voter fall-off.
Many voters vote only the Presidential line. In King County recount I saw many left the Gov race blank while voting the whole ballot (it APPEARed these were disproportionately Dem’s, but not exclusively Dem’s).
Goldy, I’m not sure I understand why you say the total includes valid votes they want to count, when both Roberts at the PI and Postman at the Times explicitly refer to 743 felons.
The number doesn’t seem as important anymore, as someone point out above, as does the establishment of principle that the numbers Rossi is putting forth as a way to get proportional analysis, are full of shit and make their request untenable to implement.
Thanks much for the transcript, by the by. Always important to get the source documents.
chew2 @ 28
”1. My understanding is that the precinct level data may be more favorable for Rossi because the Gregoire plurality in those precincts was higher than the county plurality as a whole.”
Indeed, the reduction in uncertainty that one automatically gets by going from a county-level analysis to an analysis by precincts is a statistical phenomena that arises when all else is equal. Moving to a precinct-level analysis changes the result even more if there was substantial cherry picking. This is part of the reason Katz’s precinct-level analysis showed a 95% confidence interval that excluded a Gregoire win.
Richard Pope made a beautiful analogy the other day, however. He said that if the Republicans cherry picked felons through the D precincts, then the Democrats, making the second sweep, are left to find (largely) felons in R precincts. I expect these things to largely wash out when the D and R lists are combined and an analysis done at the precinct level. I hope Goldy is right and the data are made public within a few days!
2. As I understand it using ecological inference statistical techniques it is possible in principal to estimate proportion of males who voted for rossi or gregoire, if you know the sex breakdown of voters in a precinct and the vote results for the candidates. This estimation is apparently done for race in voting rights cases to show whether blacks tend to vote for black candidates, and whites tend to vote for white candidates.
Since we know that most of the felons were male and in fact can identify each and every one of them, would it be possible to estimate the male propensity to vote for Rossi, and factor this in to an estimate of how the felons voted. I’m not sure whether the sex of the voters is known for each precinct however. But somehow the researchers are able to get race and sex data for voting rights cases.
Given some “ground truth”, it is possible to try to make adjustments for simple demographic categories like race and sex and come up with an estimate of felon voter propensities. There is a very important assumption that must go into such an analysis, however: one must assume that ex-felons and non-ex-felons, matched by demographic characteristics, vote the same. This is a most tenuous assumption with no ex-felon “ground truth” to evaluate the assumption. One could try to model the uncertainty of this assumption, but that makes it even more difficult to overturn the election. In order to get the characteristics of Gregoire and Rossi voters, one would ideally have a statewide exit poll for the last election that includes information on how people voted for governor (and their demographic characteristics).
Do you have any comments about ecological inference?
Ecological inference is commonly done in the social sciences, but it is considered second-rate compared to doing analyses that includes information about the variability within a population. In this election, the problem is severe. We have ex-felon voters (a special subgroup) but no idea how they vote compared to the non-ex-felon voters. We can roughly (and with a lot of work, precisely) identify some of the characteristics of the ex-felons (mostly white male). But this knowledge is of little use because we have no way to scientifically calibrate how a white male ex-felon, a black male ex-felon, a white female ex-felon, etc. vote. The best we can do is provide estimates, based on very crude information, of their voting patterns. An honest statistician will carefully model and keep track of the uncertainty introduced by using the crude data, and will aggregate it with other sources of uncertainty. But, all the uncertainty makes it even more difficult for Rossi to prevail with any statistical analysis.
RonK @ 33
Did your simulations take into account the fall-off (by county) of voters who left the Gov race blank?
Yes they did. I included all ballots that were not for Rossi or Gregoire as a “Other” category. Ballots were randomly drawn from the “Other” category in the proportion with which they were represented.
“There’s no way in evidence to literally satisfy Judge Bridges’ that dq’d felon votes don’t enter the case unless they voted for Governor, so I assume this envisions proportional treatment of felon voter fall-off.”
I agree. I was most surprised when I saw that Judge Bridges expected both sides to demonstrate that the ex-felon voters actually cast a vote for governor. If this standard is upheld, the Republicans probably cannot win, as it requires an affidavit from the ex-felon voters. In any case, affidavits from ex-felons are susceptible to the same “gaming” problems as asking ex-felons who they voted for. I am guessing that Judge Bridges will relax this requirement.
Goldy @ 30
That said, you’ll have to program another twist into your analysis, as the Dems’ list includes legal ballots that should have been counted, but weren’t.
No problem. That is a trivial addition. I assume we know the number of ballots but not the content of ballots, right? (I assume these are the unopened absentees found in the boxes of empty absentee envelopes).
Kentucky Derby–10 Minutes to Post Time
My picks–
High Fly 8-1
Noble Causeway 12-1
But what in the hell do I know?
Obviously I am not a very good handicapper.
Cynical @39,
We could tell that by your election contest analysis.
Cynical @33, TJ @34,
I had poorly phrased my question to the Dems, and thus got wrong information back. Indeed, the list includes 743 probable felons. They also identified 1898 mishandled provisionals, and 780 ballots in King County that weren’t counted due to errors by election officials. I have updated my post to reflect this clarification.
(Hey… I don’t think I’ve ever had the opportunity before to answer Cynical and TJ in the same comment!
DJ @37,
I assume that in the final analysis, if the Judge rules that any of these ballots should be counted, their envelopes will be opened and their actual vote recorded… so they won’t be part of any statistical analysis that might be used in the trial. But we can still use these to estimate how they might impact the final count.
I still think the GOP proportional analysis test gets thrown out under the Frye test as unreliable because the data the GOP proposes it should be used on is not reliable. If the data can be shown unreliable, for example, by showing that a significant number of illegal votes were excluded, on purpose or accidently, and the illegal votes that were missed are primarily from GOP counties, it does not matter how reliable the analysis is, because the data is flawed.
Mr. C seems to feel it is up to each party too find the illegal votes they want to prove. That is only true up to a point. The proportional analysis is supposed to convince the court who really got the most legal votes in the election, not just which party was most effective at finding illegal votes in precincts that favored the other guy. This is not like a basketball game, identify the most illegal voters and win.
Mr. C makes such a big thing out of the word “appears” in the statute, but it works both ways. If you perform the proportional analysis on a data set that does not fairly represent all the illegal votes cast, you can hardly say it “appears” one candidate or the other won. You can only show that it is possible to skew the data to get the result you want.
I think that if the court is convinced that the GOP’s data set is slanted, the proportional analysis is not going to be accepted. To do that, the dems just have to show significant (i.e. hundreds) of illegal votes were missed in GOP counties.
Goldy, (from Udate, Update)
“And there are an additional 780 absentee and provisional ballots from King County that are identified as not being counted due to error by election officials.”
There were a total of about 1000 absentee and 300 provisional ballots that were still rejected in the final count due to signature verification issues.
I am thinking that the Dems probably got this entire list, along with copies of all the ballot envelopes in question. Then they determined which ones of these 1,300 voters were most likely to have voted for Gregoire. Perhaps by actually asking them, seeing if they voted in a party primary in 2000, seeing what precinct they lived in, or by analyzing known demographic information about them.
After identifying which of these 1,300 voters were most likely to have voted for Gregoire, they may have gone to these voters with a copy of their ballot envelopes and showed it to them. The voters then recognized their ballot envelope and signed an affidavit that it was indeed the ballot envelope that they signed and mailed in (or signed and turned in at the polls).
I would imagine that very few of the mismatched signature rejections actually represented someone else (i.e. not the actual voter) signing the envelope. Conceivably, if we contacted each voter directly whose signature was rejected, we could reduce the rejection rate to nearly zero.
TJ @ 35
Thank you for a concise explanation of why the R case for guessing how felons voted is fatally flawed. I won’t be surprised if the demographics of this group turn out to favor Rossi rather than Gregoire (white, male, low level of education), but in the absence of reliable data on how felons vote that can be controlled for all other characteristics, this entire line of reasoning is bullshit.
Here’s hoping the judge has a good head for logic — if he does, the Rs are SOL.
Richard Pope @43,
Actually, no… the Supremes have already made it perfectly clear that you cannot contest subjective canvassing decisions like signature matching. The 780 uncounted ballots on the Dems list come from ballots they claim were never properly canvassed, including the 169 remaining “Phillips” ballots, which weren’t counted because officials could not find a voter registration record before the final certification deadline. In addition, there were 88 ballots marked “needs research” and 622 provisionals marked “canceled”. I’m not sure if the Dems list includes the 94 absentees found still in their outer envelopes… perhaps the R’s already included these, as they were proportionally slightly pro-Rossi.
I’m assuming that the Dems performed further research to determine if any of these were from registered voters.
that was DJ at 35, not TJ. But I agree with you, nice summary.
My apologies to DJ — thanks, torridjoe.
the fix is in at all levels. Logan a sham, Reed an embarassment, McKay watching his behind, and Supremes lacking a ethical spine
That’s right righton, everyone is in on it. Goldy, me, the newspapers and tv, your neighbors, your pets. Even chardonnay, who is really a democrat who has been assigned to make Republicans look like total whack jobs. Really good, isn’t she (or he, who really knows in cyberspace?) Better head into the lead room and put on your tin foil hat, or we will get you too.
Dear Republicans:
You lost. Got it? You lostop crying and get over it. You’re looking more childish every day.
C
Dear Republicans:
You lost. Got it? You lost. Now stop crying and get over it. You’re looking more childish every day.
C
C @ 51
Every step of the way the candidates had a choice: push it to the next level or give up. How is Rossi pursuing the contest in court any “badder” than Gregoire calling for the hand recount (and begging for all the out of state money to pay for it)? Where is the problem here? Are you saying that Rossi is doing something wrong, and Rossi voters are doing something wrong by “rooting” for him to win the contest in court?
Maybe you don’t realize that all the blogging in the world has no affect on the outcome.
If I was a Democrat I would be dreading the thought of any re-election, even in four years! She would not stand a chance in any re-election she would run for after the Bullshit lies and walking all over the citizens demonstrated in this last legislative session. The Citizens will have the last laugh of this administration!
gs,
Trick will be lack of press reporting on the facts coming out. The only way I’ve learned of the ineptitude is via SP blog. But yeah, other than the McDermott type (brain dead automatic left wing voters), the Dems overall have earned some real badwill in this.
righton,
The dems have earned badwill?
You get all your facts from snark?
The only badwill, is coming from repubs. , bent on the defamation of the democratic party. Your side has made up so many lies, and rumors, it defies comprehension.
What will be remembered after rossi gets his hat handed to him, is the lies.
I predict gregoire, will be re-elected with a large margin.
AllHat–
As Gregoire continues to increase taxes at every turn and fund every LEFTIST cause…perhaps she will get re-elected (if she is successful in getting all Washington voters hand outs and government jobs!)
AllHat is guv’mint worker, aren’t you.
Hey Head HorsesAss, Where on the “Comment on Transcript of Judge Bridges Rulings” page of SP does anyone but Micajah prematurely throw in the towel? What is on Stefan’s web site is an honest attempt to look at all sides of the judges ruling issue. You and the gang at HA don’t even attempt to look at all sides. You all bloviate, riducule, name call, and stigmatize anyone who has the guts to write on this blog an opposing viewpoint.
I guess that’s why you are looking at the south bound end of a north bound donkey everyday!!!
Pudster
hitting the bong early this morning, eh pud? Fire one up for mom, willya?
snark is full of shit, generally.
Most of his premature ejaccusations, have gone by the wayside, as will rossi.
It doesn’t take guts to post ridiculous accusations, it takes guts to admit you were wrong. Has he admitted he was wrong, and I missed it?
It doesn’t take guts to keep pitching the same b.s., knowing its wrong, just to stir the troll base.
It doesn’t take guts to ban all i.p. addresses, of posters with differing opinions.
You see anyone banned here?
How much is the tax hike going to cost you cynical?
the average person will pay less than $100/year from the gas tax.
Puddinghead @57,
Do you always comment on things before you read them? There was quite an atypically frank discussion going on the comment thread between Micajah, and a number of other people.
And again… don’t you ever feel just a tiny bit stupid constantly accusing me of shutting down debate, when unlike almost every other political blog, I clearly let all comers post whatever they want… even assholes like you? Even Cynical gives me grudging respect on that point.
righton @54,
Yeah… well… if you’re getting all your news from the Snark, that explains a lot.
Let me ask you a question… Snark clearly misled you into believing Rossi actually had a case… what else do think he might have misled you about?
AllHat@59–
“the average person will pay less than $100/yr. from the gas tax”.
And this is on top of how many other ONLY $100/yr or $50/yr or $20/yr. taxes AllHat??????
The problem with Guv’mint workers and Guv’mint pensioners like AllHat is they have never risked capital or employed anyone or woke up in the morning not knowing if they were going to make a buck or not. The AllHat F***ing Loser brigade are scumsuckers and bottomfeeders that get paid JUST FOR SHOWING UP!!!!!!
They do nothing of real value for society and suck away tax dollars from where they are really intended.
If I hear one more bottomfeeder like AllHat tell me it’s only another $100/yr., I think I will…………._______________.
AllHat==Pathetic Bottomfeeding Loser!
CynicalCryBabyWhiner@IDontWannaPayForNothing,
I dont work for the government, you pencil dick.
I dont hear you pissing and moaning about what rossi is costing the state with this stupid lawsuit.
I dont see you pissing and moaning about how much bush’s tax cuts will cost everyone in this state.
I dont see you pissing and moaning about how much of a “TAX” increase, you got from big oil.
Loud mouth, bullshit does something FOR society?
If so, you are to be highly commended.
AllHat
$9 billion divided by the popultaion of this state is a whole lot more than implied by your “$100 per year”. Don’t forget the population includes kids, elderly and poor people. So tell me again how little this pack of tax hikes is going to cost us but pull your head out of your ass first.
“what else do think he might have misled you about?”
Comment by Goldy— 5/8/05 @ 10:49 am
Goldy you’ve admitted that you deliberately post hype to fire up the lefty base. On second thought I shouldn’t say you “fire up” I should say that you “mislead”. All your articles about HB 1515 are a prime example of this tendency.
Now you’re throwing stones the other direction?
AllHat@62–
So how many years did you have a guv’mint job dude??
How do you earn a living???
Trustafarian?? Go to the mailbox and get yer check.
Get yer allowance from Mommy and Granny.
You are a lying sack of sh** and you know it.
No real business owner would spew the crap you do.
I am just tired of funding useless Guv’mint pinheads like you.
Gimmee infrastructure, military and public safety….I’ll gladly pay for that. The Alaskan Way Viaduct “wet-dream” for the People’s Socialist Republic of Seattle??? Screw that. You idiots have pissed away hundreds of millions planning and replanning. Now you have nothing left to do the project. Charge a toll and good luck.
Every step of the way the candidates had a choice: push it to the next level or give up. How is Rossi pursuing the contest in court any “badder” than Gregoire calling for the hand recount (and begging for all the out of state money to pay for it)? Where is the problem here? Are you saying that Rossi is doing something wrong, and Rossi voters are doing something wrong by “rooting” for him to win the contest in court?
The Republicans are just doing the whining thing to keep their wingnut troops angry. You guys specialize in having enemies.
YourZippersDown@63,
GAS TAX.
AllHat@62-
“I don’t work for the government”
That is undoubtedly the truest thing you have ever said.
Although you do collect your paycheck & benefits, you are ENTITLED to that somehow without work that benefits we the people. Nice scam dude.
CynicalDicklessBiawHack@LosingSteam,
How nice of you to gloss over your meager pittance from the biaw coffers.
Pencil dick, was the truest thing i ever said about you.
Quit trying to dramatize.
If you can swallow all the b.s. from the right wingers, you can swallow a little tax increase.
Shrubby has buried us in tax increases for the next 2 generations, and if thats ok, then you can fuck off about the budget.
Goldy @ 60.
a) Well before Shark I learned you can’t rely on our twin socialist rags, so he’s about all i got (rick anderson is good, even sometimes the stranger). Suits on the radio, sometimes the other local right wingers.
b) Not sure how you figured out (erroneously) i or others bought into some thesis he issued and was wrong about…(i didn’t buy in, not sure he said it, today not sure if he did say it, that he’s wrong)
Allhat…
Why the profanity? Sometimes serves a role, but you overuse it.
righton@69,
you’re right, if shrubby’s burying us is ok, cynical can ignore the wa. budget.
When responding to cynical, i feel the keyboard is a weapon.
allhat @ 61
That $100 per person you’re tossing around is pure BS. The gas tax bill is going to raise $9.5 billion. How do you get $100 per “average driver” per year when only 6 million people live in WA and a portion of those don’t drive? Do the math knucklehead.
slowly now….15000 miles a year, 20 miles per gallon, .095 cents a gallon, $71.25.
I did the math, knucklehead.
allhat
How do you get $9.5 billion out of that? You can’t.
i would draw your attention, to the entire budget, and the time frame involved.
The gas tax, is .095 cents a gallon.
AllHat@72–
The keyboard is your “weapon”!!
Wow..not too many LEFTIST like you use the word “WEAPON”.
It’s politically incorrect….violence ya know.
Perhaps you need to get into the Lotus position and hummmmmmmmmmm. Visit http://www.shantimai.com
Have you lefties been out talking to non blogger lefties lately who voted for Gregoire. I was not shocked but pleasantly surprised today to hear from two of my friends who are democrats who voted for Gregoire, who now and their families are clearly against her current direction. One says She is trying to make this another California, and is puttin ght democratic party of this state at risk with a bad name, the other told me they have given up support for her after her disregard for I601 taxing limits. I for once just shut up and listened thankfully to both of the people who are diving off the sinking Tax and spend ship. She is imploding your party!
Did you know, over 1/2 the people in the country that voted for shrub, that still believe iraq has wmd?
People are funny.
It helps to to sway people, when you have the wacko right, claiming fraud, corruption, and lying, regardless of the fact these things are just mud flinging.
GS@78: I live in a very conservative area, pre-dominately Republican (affluent retirees). One of the most prominent, respected, member of our community, a conspicuous and dedicated Republican, had this to say after Gregoire gave a speech to our local Chamber of Commerce: “I voted for Dino Rossi. But Gregoire is not a bad choice for Governor. I admire her intelligence and courage” and went on to say that he thought she was doing a good job as Governor. This person is a CEO of a larage organization (energy business)- well respected and well known.
Smart, thinking, people, Yes, even Republicans, are impressed with Governor Gregoire and are saying so, out loud and in public.
Get used to it. Marilyn
GS @ 78
Your anecdotal examples don’t count for much (in the same way the anecdotal evidence that most felons voted for Rossi doesn’t go very far). I’ve not seen any polls on Gregoire since the end of the session.
I have, on the other hand, seen the miserable ratings that Shrub is polling at. He is doing about as bad as any time since he has been in office, tying his miserable ratings from May of last year. Check ‘em out at http://www.pollkatz.homestead.com/
GS@78: I realize that anecdotal examples can’t compete with polls in identifying trends. However, this is a little past mere anecdote – it is not so much what was said, but who said it. The individual who made that statement is one of our most conspicuous and respected citizens, quite active in Republican party circles as well – the local press picked it up – as they would have reported anything else this individual said. There isn’t much out there in concrete complaints about Governor Gregoire anyway, even for the most rabid conservatives – it’s good thing they now have the gas tax to alleviate some that pressure on their spleen.
I too am a fan of pollkatz, Daily Kos, dKospedia, etc.
Marilyn
Hell, I think the gas tax ought to be about $2 a gallon.
C@83: $2.00 a gallon gas tax? If demand stays steady at the current price, we ultimately may see that. Were you thinking at the state level? or at the federal level?
Marilyn
Mr. Cynical it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year it’s only $100 per year hope this helps you in your _____________
Marilyn; ok so you claim Jim Ellis or kin likes Gregroire? I think they all are related to Sam Reed, and/or look alikes of lib ex gov Booth Gardner (talk about a good old boy netowrk)