More errors in GOP felon-voter list

A quick link to David Postman’s update on the GOP felon-voter list in the Seattle Times. As has been reported previously, the list includes hundreds of names of people convicted as juveniles, who did not have their voting rights revoked. Now as the various counties attempt to slog through the list, more and more errors are being uncovered.

“It’s really inconsistent information. But 1,100 felons voting makes a great headline,” said Yakima County Auditor Corky Mattingly.

She said that earlier this year, the county reviewed a list of 15 alleged felon voters compiled by Rossi backers but not submitted to the court. She said four of those 15 appeared to have voted improperly.

Republicans now say 31 felons voted illegally in Yakima County. Mattingly said she’s seen the list of names, but there are no birth dates or criminal case numbers, which makes the list difficult to research.

A 73 percent error rate? Is that an anomaly? Apparently not:

Whatcom appears to be one of the few counties that has completed a thorough review of the list.

According to the Republicans, the county had 13 illegal votes cast by felons. But county Auditor Shirley Forslof said a review by the court clerk found many errors:

Four of the voters in question were convicted of felonies but had their voting rights restored. One was entitled to vote because the conviction was for a gross misdemeanor, not a felony. One was free to vote because the felony case had been dismissed. Three on the list didn’t vote in the November election, contrary to what the Republicans had said.

The remaining four voters, Forslof said, appear to have voted illegally, and their names have been forwarded to the county prosecutor.

I know, I know… whether the final number is 1,100 or 400 or something in between, hundreds more felons voted illegally than Christine Gregoire’s margin of victory, and that, say Rossi supporters, should be enough to set aside the election. Yeah, well, perhaps in their minds, but not in a court of law, where Rossi will actually have to prove that these illegal votes and other errors changed the outcome. As GOP allegations consistently prove to be less reliable than the election itself, even their proportional analysis strategy looks unlikely to provide the desired result.

The Rossi camp went into the contest expecting to find thousands of illegal votes and irregularities, if not evidence of actual organized fraud and corruption. Whether their optimism (or was it pessimism?) was due to faulty initial analysis or a genuine faith in the incompetence and immorality of Democrats — or just plain blind hope — their investigations simply haven’t panned out. The normally scandal hungry MSM, that was so eager to jump on earlier allegations, has become rightfully skeptical of charges of a “stolen” election… and even right-wing talk radio seems to be growing tired of the twice daily spreadsheets from that guy on the other blog.

The truth is, we had a run-of-the-mill imperfect election, with the normal assortment of randomly distributed errors. Under those circumstances it simply was not possible to confidently determine the winner of such an extraordinarily close election. Some people understandably find such uncertainty reason enough to call for a new election. But unfortunately for them — and Dino Rossi — our election statutes do not.

Comments

  1. 1

    Dubyasux spews:

    If it turns out the majority of illegal felon votes went to Rossi, as the anecdotal evidence indicates, then it’s to Rossi’s advantage if that list shrinks. Who knows, he might even win if the GOP illegal-felon list goes into negative numbers. For example, if BIAW’s corrected list has minus 1,100 illegal felons on it, and 60% of them voted for Rossi, then Rossi got minus 660 illegal votes and Gregoire got minus 440 illegal votes, which is a net difference of minus 220 illegal votes, and when you add this to Rossi’s and Gregoire’s vote totals Rossi would win by 91 votes. Yes, I finally see the light at the end of the tunnel for Rossi; there exists, after all, a scenario under which he has a chance to win.

  2. 2

    theREALanonymous1 spews:

    there exists, after all, a scenario under which he has a chance to win. -Comment by Dubyasux— 3/27/05 @ 12:34 pm

    Which is why all this conjecture, hypothesizing, second guessing, predicting, theorizing, presuming and speculating is a monumental waste of time.

    I suppose it has some value as entertaining blogging though… not unlike reading your daily horoscope.

    It will be what it will be and when the courts decide it will be.

  3. 3

    Richard Pope spews:

    I think the proportional analysis method is the only one that makes sense to deal with illegal or otherwise improper votes, when these votes are not the result of organized fraud.

    For example, in many states, the laws expressly provide for a similar solution when the number of ballots cast in a given precinct exceeds the number of people actually voting. In Wisconsin, they will randomly draw that number of ballots out of the precinct ballot box until the number of ballots gets down to the number of legal voters. I am sure other states have similar laws.

    When there is no organized fraud, I can’t think of any other appropriate solution for excess ballots. It would be ridiculous to ask improper voters who they voted for. They could say they voted for the other candidate, so that the candidate they really prefer gets a boost by another vote.

    We also have several hundred excess ballots in King County that can’t be attributed to specific voters. There may be 780 or so excess absentee ballots, and maybe a few hundred extra poll place ballots (not including the directly inserted provisional ballots where the voter was actually qualified).

    Probably the election contest will end up with every single page of the poll book sign in sheets and every single absentee ballot envelopes being double checked for King County, to make sure there really are more ballots than voters and make certain as to exactly what this number is.

    In the end, either proportional analysis, or a computer program to randomly reduce vote totals in each precinct (based on the vote in each precinct) will probably be used to deal with the illegal voters and phantom ballots.

    In the meantime, the GOP needs to double check the felon voter list. They should have a file for each of those names, which should include (1) a certified copy of the judgment and sentence for an actual felony conviction and (2) a certified copy of the court docket sheet showing that no order restoring voting rights has been entered. The illegal felon voter charge cannot be proved in court without these two documents, so the GOP should at least have these two documents in its possession before accusing these people of voting illegally.

  4. 4

    Janet S spews:

    What I found astounding was the level of derision from several of the county auditors. I’m still puzzled – if the county auditors are responsible for maintaining the integrity of the voter lists, but refuse to do it, who is going to do it? It is nonsense to say that they have no legal ability to remove names. Of course they do, and they can refer those voting illegally to the prosecutor.

    Admit it, if this case wasn’t going forward, this election and all its problems would now be forgotten. I still haven’t heard a good explanation for why there are so many more votes than voters. Someone fed lots of ballots into the machines, either by mistake or on purpose. Without this case, no one would care, particularly the majority of the readers of this blog.

  5. 5

    Dubyasux spews:

    Richard @ 3

    Proportional analysis is, when you get right down to it, conjecture and speculation. Which is why Judge Bridges won’t go for it. Courts don’t decide legal issues based on conjecture and speculation. They require proof.

    Janet @ 4

    Only in your imagination, Janet.

  6. 6

    Mrs. Cynical spews:

    Richard Pope @ 3

    > In the end, either proportional analysis, or a computer
    > program to randomly reduce vote totals in each precinct (based
    > on the vote in each precinct) will probably be used to deal
    > with the illegal voters and phantom ballots.

    Naaa. . . neither solution works, at least for the felon votes. The felon voters are not even close to a random sample from the voting population. They are disproportionately lower socio-economic status and male. Such an adjustment would disproportionately disenfranchise women, individuals in high income brackets, Asian Americans, etc.

    Of course, we shouldn’t overlook the other reason it will not work: the WA statutes make no provisions for either solution.

  7. 8

    Erik spews:

    and that, say Rossi supporters, should be enough to set aside the election. Yeah, well, perhaps in their minds

    Yep. And under their theory, no election could ever be won when the margin of victory was less than the error rate, which is likely at least 3000 statewide.

    However, the election statutes demand a winner.

    Furthermore, even if there was another election and people voted the same way, the difference between the candidates would still be the same. Therefore, the re-vote theories are unworkable.

  8. 9

    Diggindude spews:

    Maybe this new republican “tupperware” party idea will help.
    1000 points,(i mean “sparks”) of light!!
    Maybe rossi and vance can ask the felon voters to sign in, for identification purposes.

  9. 10

    zapporo spews:

    ahem…..The other avenue seems far too likely. Dozens and dozens of botched vote counting activities by King County that far surpass 129 votes. Ballots appearing from nowhere, counts that are not consistent, provisionals counted without verification, and the felons, dead people and double voters on top of that. I think that, even in the smallest, greenest, grinchiest liberal heart, we all know that Washington deserves a new election. It will be very interesting to watch Mr. Logan’s deposition, coming soon to a court tv near you.

  10. 12

    zip spews:

    Don @ 1

    After you read this:

    http://www.asanet.org/media/felons.html

    you may want to re-review your anecdotes that indicate the majority of felons voted for Rossi. I have seen this claim trotted out time after time with nobody ever offering anything more than vague references to anecdotes. How about some reason why the Washington felons would differ so much from the national pattern?

  11. 13

    Goldy spews:

    Zip… the point about claiming anecdotally that the majority of felons voted for Rossi, is that it is just as ridiculous as the claims that a majority voted for Gregoire. I know… “Democrats are more likely to cheat than Republicans.” You may truly believe that. But statistically, that assertion is a load of crap.

  12. 14

    chew2 spews:

    ZIP @12

    Studies that estimate how felons voted, including the one you cited which suggest a higher democratic vote by felons, are not based on any surveys about how felons actually voted, but are only statistical estimates based on the what we know about the felon population, such as age, education, sex, race, income, propensity to vote, etc.

    There are far fewer black felons in Washington (who tend to vote democratic) than in the national sample. Since we should know the age and sex and possibly race of the felons who voted, we can make an estimate of how they voted based on those demographic factors. You’d also have to factor in the fact that the democrat was a woman, and the republican was a man.

    I’ll bet it will split 50/50 or tilt slightly to Rossi. Even using that level of detail will probably be found by the Supreme Court to be too speculative to comply with the Washington statutes governing the election contest.

  13. 15

    dj spews:

    zip @ 12.

    We’ve covered this territory before. The summary you cite is based on a research article by Uggen and Monza published in the American Sociological Review, volume 67, number 6 page 777-804 in 2002.

    While the inference probably holds nationwide—felons overall tend to be democratic—the inference is largely based on the proportion of black felon (potential) voters, and assuming black felons would vote like black non-felons who, as a group, tend to vote Democratic.

    Nationally, the proportion of black felons is about 50% (based on Department of Justice Statistics, for Jan 1 2001). In Washington, however, only about 15% of felons are black. In other words, Washington has a much higher proportion of white felons compared to the nation as a whole. The vast majority of our felons are white males, who are more likely to vote Republican.

    Take-home message: don’t take national statistics and try to apply them to regional/local politics. The “felon-voter effect” does not appear to be beneficial to Democrats in Washington. It is probable that felon votes in Washington (if they were allowed) would favor Republicans. The anecdotal evidence from confessed illegal felon voters in the last election is consistent with this idea.

  14. 16

    Dubyasux spews:

    zip @ 12

    The only thing I would add to what the others said is, you’re relying on pure guesswork, whereas the “anecdotal evidence” I referred to is that the few ex-felon voters who have talked to reporters overwhelmingly say they voted for Rossi.

    Is this a statistically valid sample? No. Do we know for sure they’re even telling the truth? No. But it’s the best information we have right now, and it’s more than nothing.

    Proportional analysis, national studies (such as you cited), and the wishful thinking of Rossi partisans are pure conjecture.

    Conjecture is not anecdotal evidence or any other kind of evidence, it’s nothing.

  15. 17

    Richard Pope spews:

    Why is everyone knocking the felon voter report from the American Sociological Association? That group is not exactly a Republican think tank. It has a membership of 14,000, most of whom have PhD’s in sociology and are professors of that discipline in colleges and universities throughout the USA.

    http://www.asanet.org/about/about.html

    Most sociologists tend to be liberal and Democrat leaning. Nothing conspiratorial about that — it is just they way that most social science professors tend to be. I would bet that Uggen, of the University of Minnesota, and Jeff Manza, of Northwestern University, the two sociologists who wrote the ASA study, favor Democrats in the personal voting as well.

    So let’s just suppose that we have 800 felons improperly voting in King County. If their precincts, on average, have the same voting patterns as the whole of King County, then an 18% margin to Gregoire takes away 144 net votes on the proportional analysis method, and can be expected to take away an average 144 votes if a random ballot or vote reduction lottery is conducted. This wipes out Gregoire’s victory margin and could even make Rossi governor outright.

    Moreover, the scientific research (including ASA) indicates that convicted felons are more likely to vote Democrat than average people. The ASA factors that influence this the most are (1) higher than average percentages of blacks and Latinos among felons, (2) higher than average percentages of high school dropouts among felons, and (3) higher than average percentages of lower income people among felons.

    It is true that felons also have a higher percentage of men than average. This is a slight push towards Republican. However, the gender gap spread is a much smaller influence than the racial/ethnic spread, the education spread (high school dropout v. graduate), and the income spread (bottom income vs. the rest). So the ASA study has about 70% of felons voting Democrat, vs. about 50% for the population as a whole.

    DJ is right that 50% of felons nationally are black, vs. 15% being black in Washington. However, this does not mean that felons in Washington are more likely to vote Republican than Democrat. Washington only has about 3% black population. So a subset with 15% black, such as convicted felons, will still have more blacks (a group which generally votes 85% Democrat) than the state population as a whole — and should be more Democrat than the state’s voter population as well.

    Bottom line is that there is no scientific study or other admissible evidence that would show that convicted felons in Washington are more likely to vote Republican than are other people living in their same neighborhood. And all of the studies on the matter (nationwide and in other states) point in the exact opposite direction.

    I don’t think Judge Bridges is going to allow these studies into evidence. Even if the scientific reliability were to be established of a particular estimation technique, it would be offensive to public policy. This would involve determining how each illegal felon voted based upon his or her race, education, income, and gender. That just isn’t going to happen.

    The only reliable and neutral method is to either use a proportional analysis, based on how all the people in the same precinct voted, or to randomly eliminate votes from a given precinct using these same percentages.

  16. 18

    Richard Pope spews:

    Oh — and by the way — sociology professors are entitled to make nominations for the Nobel Peace Prize, just like members of Congress may do :)

    However, the overwhelming majority of sociology professors are intelligent enough not to make nominations for the Noble Peace Prize in Medicine. Only dumb members of Congress are unqualified enough to make such a bone-headed move.

  17. 19

    dj spews:

    Richard @ 17

    I truly do appreciate some of your posts, however, I think this post is “off” on a number of points. First, nobody is dissing the ASA anywhere that I can see. The ASA is not responsible for the content of the Uggen and Monza paper—they only summarized it.

    Secondly, several posters have pointed out shortcomings of the Uggen and Monza study. But, if you read the study, Uggen and Monza also point out these limitations. They had to make imperfect inferences [using national age/ethnicity/education, etc. statistics to approximate patterns for (ex)felons] because a better, more direct source of data does not exist.

    You wrote:
    “DJ is right that 50% of felons nationally are black, vs. 15% being black in Washington. However, this does not mean that felons in Washington are more likely to vote Republican than Democrat.”

    It most certainly does! The inference about the percent Democratic-leaning felons (nationally) is largely a function of 50% blacks in the national felon population. If Washington has only 15% felons, then the proportion of Democratic-leaning felons (in WA) is reduced to something much closer to the white male figure.

    You wrote:
    “Bottom line is that there is no scientific study or other admissible evidence that would show that convicted felons in Washington are more likely to vote Republican than are other people living in their same neighborhood.”

    I agree! But, at the same time there are no scientific studies that would show that (ex)felons in Washington are more likely to vote Democratic, either! There is far too much error in just adjusting for black/white differences. We would also need to consider Washington-specific differences in age, sex, education, etc. in order to even get close. But, this is all still based on the questionable assumption that (ex)felons would vote the same as non-felon voters when matched for ethnicity, age, sex, education, etc. A social scientist may accept such questionable assumptions for asking “what if” questions for the sake of intellectual exploration. No social scientist would accept the validity of applying the results to statistical correction of an election!

    Finally, you wrote:
    “The only reliable and neutral method is to either use a proportional analysis, based on how all the people in the same precinct voted, or to randomly eliminate votes from a given precinct using these same percentages.”

    I am afraid I cannot agree. I think Mrs. Cynical @ 6 hit the nail smack on the head: “The felon voters are not even close to a random sample from the voting population. They are disproportionately lower socio-economic status and male. Such an adjustment would disproportionately disenfranchise women, individuals in high income brackets, Asian Americans, etc.” (Woah. . . You go girl!).

    As an aside, Richard, after the 2000 census was taken, the census bureau wanted to statistically correct some urban census tracts for underrepresentation of minorities. It seems that some minorities are distrustful of census takers and avoid filling out the form. The resulting underenumeration had tangible negative consequences: it meant that fewer dollars would be funneled into these areas for social services. But, also, there were implications about how Representatives were allocated and how political boundaries would be drawn. Demographers could demonstrate that they could closely approximate the amount of underenumeration and thereby correct areas of high underenumeration. Yet the courts rejected statistical adjustment of census data–the stakes were too high and there was some uncertainty involved. My impression is that courts are exceedingly reluctant to apply statistical adjustments, particularly when the uncertainty involved can result in new harm being done to innocent parties (i.e. women who are “disproportionally disenfranchised” by your proportional analysis or random elimination scheme).

  18. 20

    dj spews:

    Richard,
    A correction to my previous post. You said, “However, this does not mean that felons in Washington are more likely to vote Republican than Democrat.” I read your statement incorrectly, sorry. My point is that a Washington (ex)Felon is much more likely to vote Republican than an (ex)Felon in general. After playing with the numbers for a few munites I agree completely with Chew2 @ 14 that the (ex)Felon voters in Washington probably split about 50/50. But that figure has lots of biases and errors. I could see it going 15% in either direction. The only “ground truth” we have is the anecdotal “confessions”, and that evidence has its own potential problems (See Dubyasux @ 16).

  19. 21

    Mr. Cynical spews:

    There is much more to this election contest than ust the felon issue. Illegal provisional ballots–proportional analysis definitely applies. That will reduce the 129-vote margin substantially….it may even eliminate it when combined with double votes, forgeries and some other goodies.

    We haven’t even begun to consider the impact of KingCo procedural errors and negligence with the poll book reconciliations and absentees.

    The felons are important but one of many issues.

  20. 22

    jpgee spews:

    WOW, MRs. C, yo ulet the idiot of f his cage again, it has been so nice here with just about 5 trolls, now the ‘troll of the century’ shows up again….must be looking for boyfriends

  21. 23

    jpgee spews:

    Not availble idiot, have 5 dogs, 5 chickens, 3 squirrells, 6 birds and 16 fish, and each and everyone of them has more sense than you

  22. 25

    chew2 spews:

    Pope,

    You say: “The only reliable and neutral method is to either use a proportional analysis, based on how all the people in the same precinct voted, or to randomly eliminate votes from a given precinct using these same percentages.”

    I have to agree with the others that this claim is invalid. I believe, someone posted a link in a prior thread to the statistical literature which explains why your apportionment method is scientifically bogus. The only reason one can claim that a felon likely cast their vote as their precinct as a whole is because they share some common attributes with those in the precinct. If they do not there is no basis to say they would have voted the same. Why select the precinct? Why not the county, why not the region, Why not the state? What makes it any more likely that any of those geographic groupings are a more accurate indication of how a felon would vote? As a statistical matter your proposal is far more innaccurate than attempting to estimate the felon vote based on their known demographic attributes. Even the latter is too uncertain to support an election challenge.

    The problems in estimating how someone cast their vote lends support to the statement by the Supreme Court in the old Howell case that if you didn’t know for whom the illegal vote was cast you should presume it was a valid vote.

  23. 26

    Richard Pope spews:

    DJ @ 19 & 20

    Simply because the felon population in Washington is “whiter” than the national felon population as a whole — this does not mean that Washington felons are less likely to vote Democrat than the national felon population. Nor does it mean than Washington felons are more likely to vote Republican than Washington non-felons (i.e. general population).

    White people in Washington are much more likely to vote Democrat than white people nationally. The 2004 CNN exit poll had white people in Washington voting 52% Kerry to 46% Bush. Whereas, nationally the figures for white people were 58% Bush to 41% Kerry.

    If you took a nationwide sample of felons (which should start out around 50% black), and then eliminated blacks from the sample until the sample was only 15% black — then maybe the nationwide sample would be close to 50-50 between Democrat and Republican voting.

    However, if you took a random sample of Washington voters (which should start out around 49% Gregoire and 49% Rossi, and should also contain about 3% black), and then added randomly selected black voters to the sample until it was 15% black — then we should expect the sample to be much more heavily weighted towards Gregoire, instead of being nearly balanced as the original sample was.

    But my main point is the lack of valid scientific data to show that Washington felons are likely to vote differently than valid voters in the same precinct. In the absence of such data (and in the absence of organized fraud), then proportional analysis or random vote/ballot reduction at the precinct level is appropriate to deal with this problem.

  24. 27

    chew2 spews:

    Pope,

    One other question. The rule you propose would be a judge made rule, since the Washington statutes doesn’t provide for such an apportionment. I can see some arguments for the Wisconsin rule, but it was enacted by the legislature.

    Here it would be judical lawmaking of the broadest kind, something that conservatives supposedly frown on. Here the judge would consider various methods of estimation, and choose a less accurate method based on pure public policy grounds of not wishing to take into account race, sex etc. Would you support that?

  25. 28

    Wayne spews:

    Mr. C.:

    Unless the republicans prove the improper provisionals and the alleged “ballots without voters” are illegal votes, they won’t factor into the analysis. The court should not presume such a vote is illegal, the Republicans have to prove it. Your analysis presumes they are illegal. The court is not likely to accept the presumption. Therefore, the push to find as many felon voters in King County as possible. However, if the Democrats are smart, they are combing eastern Washington for felons from red counties. If they get anywhere near the same number of felons from Rossi counties, the proportional analysis will not resolve anything. In fact, what the R’s appear to be doing is suspiciously like what Gore was accused of doing in 2000, cherrypicking the counties.

    Richard @ 17:

    Proportional analysis by precinct breaks down when you have only a few votes per precinct. If you have 50 precincts, each voting for Gregoire by between 51 and 60%, and one felon per precinct, the vote in each precinct would be subtracted from Gregoire, taking away 50 votes. Yet it is obvious not all 50 would have gone to her. You also can’t take onvious demographic differences into account. If we assume that men were more likely to support Rossi and Gregoire got more votes from women, the predominately male felon vote in King County probably was not tilted as much to Gregoire as the county overall. Yet we can’t say exactly what those percentages would be, only guess from polling data. I would argue any proportional analysis would be inherently inaccurate if it does not take gender factors into account. I can’t imagine a court accepting that kind of uncertainty as the basis for overturning an election. Unless proportional analysis results in a huge shift, so that it is obvious the election result was incorrect, I don’t think you will see the election overturned.

  26. 29

    Goldy spews:

    I agree with Wayne… the huge problem with proportional analysis (from a pragmatic standpoint) is that the numbers are too small to support it. If there were thousands of illegal votes in KC, I could see a judge concluding that it was likely to have favored Gregoire. But proportional analysis is too rough to be applied to such small samples.

    The law says Rossi has to prove he one, and under these circumstances, I don’t think proportional analysis is proof.

  27. 30

    dj spews:

    Mr. C @ 21,
    This is a point worth discussing. . . a little bit.

    “Illegal provisional ballots–proportional analysis definitely applies. That will reduce the 129-vote margin substantially”

    The same arguments apply as for the felons—how representative are (for example) provisional ballots of the overall voting population? Can we even ascertain whether provisional ballots are representative of improperly-cast provisional ballots?

    If proportional analysis is even allowed (and I doubt it will be), a Judge is going to be extremely conservative in throwing out a certified election.

    In an experimental setting (i.e. under perfect conditions), a judge could apply proportional analysis. Suppose King County was the only county in which there were errors, and the erronious ballots were exactly representative of the 59%/41% D/R split. It would require 725 ballots proportionally removed at 59%/41% from the D and R columns in order to produce a one-vote victory for the Rs. In this case, the judge has a 50% error rate. That is, if we took Richard’s method of randomly drawing 725 ballots from the total pool, and repeated the exercise hundreds or thousands of times, we would find that we overturn the election 50% of the time. If follows that 1/2 the time the _wrong_ call is being made. Overturning a certified election by making the wrong call with a probability of 0.5 would never fly.

    What level of certainty would a judge need? Scientists typically use 95% certainty to accept a finding as “significant”. But, in science the stakes are not so high. Would a judge be content to accept that he or she has a 1 in 20 chance of mistakenly overturning a certified election? I doubt it. What about 99% certainty? If many elections were to be challenged this way, could society (or the judge) accept that 1 in 100 elections were mistakenly overturned? Perhaps 99.9% certainty is called for? Who knows. But, we can at least compute the number of statistically cancelled ballots needed for each level of certainty.

    For 99% confidence in making the right decision, a judge would have to statistically cancel 800 ballots in KC. To be 99.9% certain the judge would have to statistically cancel 1010 ballots.

    I mentioned that this was an ideal experiment. In the real world there are other reasons why the proportional method would have much greater uncertainty, requiring many more statistically canceled ballots. Here are a few:

    1. We assumed that other counties had no errors. Obviously, we would have to do an analysis for each county in this way and aggregate up to a state-level election outcome.

    2. Furthermore, we would have to give all counties the same level of scrutiny as KC. This would be necessarty to ensure that each erroneous ballots in the entire state has the same chance of being uncovered as an erroneous ballot in KC.

    3. We assumed that the invalid ballots are represented by the voting behavior of the valid voting population within a county. This is an extremely dubious assumption. Obviously, we could do a little better by doing the analysis at the precinct level Even then, how can you be sure the “stuffed” provisionals are properly represented by the valid vote proportions?

    There are other reasons as well, but I don’t think we can realistically get past #2 and #3.

    Mr. C., I know you _wish_ that the judge would overturn the election. But, do you realistically believe a judge would risk improperly overturning a certified election using a statistical method with flawed underlying assumption? Rossi’s only realistic hope is to prove a conspiracy that changed the election by vote tampering.

  28. 31

    spews:

    dj @ 19 sed:
    “But, this is all still based on the questionable assumption that (ex)felons would vote the same as non-felon voters when matched for ethnicity, age, sex, education, etc. A social scientist may accept such questionable assumptions for asking “what if” questions for the sake of intellectual exploration. No social scientist would accept the validity of applying the results to statistical correction of an election!”

    Bingo. If you read the ASA paper, you’ll discover that the researchers admit there are no data on felon voting patterns; what they have done is made assumptions based on their other characteristics (race, age, income, education, et al). What they don’t admit is that their process entirely ignores the very real possibility that felons vote in a manner that is distinct based on their behavior, not their attributes. That is to say, a black male under 30 with no high school diploma might vote one way, but one who is a felon might not vote differently based on their experiences in the justice system. That young, uneducated black male may in fact be more likely to vote in the same way as a 45 year old white female with some college, who is also a felon.

    There is a further confound involving the stage of “felonhood” a person is in. Would felons who have completed their penalty, are off parole and probation, and have made restitution, vote the same way either non-felons or non-”completed” felons would? We have no idea. It would appear that many of the actual felon voters in WA are those who believed they had become eligible (and in fact were eligible), but had not undertaken the paperwork required to make it so.

    Pope @ 26 sed:
    “But my main point is the lack of valid scientific data to show that Washington felons are likely to vote differently than valid voters in the same precinct.”

    Uh, really? If that’s true, then the ASA paper is, in your judgement, not valid scientific data. And so why are we talking about it?

  29. 32

    spews:

    dj @ 30

    Your analysis makes sense, but let’s not forget that only a small percentage of provisional votes have turned out not to be cast by valid voters. King reports a validation rate of over 80%. So if we hypothesize 700 provisional ballots misfed (which is Sharkansky’s latest number I believe), 560 of those would be legal ballots, cast in a procedurally incorrect manner. Using a broad proportion of 60/40 for Gregoire, that’s all of 28 ballots.

  30. 33

    Chee spews:

    When in doubt, punt. Rossi chose to punt. The Rossi Rouse is more apt to get nowhere in spite of all the revelling that has went on. Regardless of error, reform will continue to be put in place. Majority of people are tired of the Rossi issue. Most people have tuned Rossi out long ago, moved on and Rossi needs to unpack his bags and give up moving in the governor’s mansion, it is already occupied and more apt than not it will remain so till term ends.

  31. 34

    Jeff B. spews:

    What Goldy is saying here is that this time the Democrats got away with murder, and due to the way our laws don’t allow someone to be tried twice for the same crime, that’s just too bad. But, like we all know OJ did it, we also all know that this election was too close to call.

    It’s all morally relative for Goldy and others like him. As long as Gregoire is in office and there’s nothing anyone can do, Goldy is happy and the ends have justified the means.

    Forget about truth and justice.

    We’ll see what happens in court. What Goldy fails to acknowledge is that the election may indeed simply be thrown out on the grounds that nothing adds up or is truly explainable. It’s truly not deterministic given the poor records that King County kept. Judge Bridges can certainly throw out this election, and he may well. The question is whether our activist liberal Supreme Court will let that decision stand.

    For now, Gregoire and the Democrats are certainly having fun with a legislative orgy that would make any public employee union boss smile. Enjoy it while it lasts.

  32. 35

    spews:

    Jeff @ 34
    “What Goldy fails to acknowledge is that the election may indeed simply be thrown out on the grounds that nothing adds up or is truly explainable.”

    Perhaps he’s not acknowledging it, because it’s a legal impossibility for him to do so? “Not adding up” is not sufficient grounds for setting aside anything.

  33. 36

    angryvoter spews:

    Torrid are you serious? If the numbers dont add up it isnt an issue? I hope you dont work at my bank.

  34. 37

    spews:

    angryvoter @ 36:

    By your tally, we should be throwing out pretty much every election in the nation, since essentially none of them actually balance up.

    You’re equating bookeeping with vote counting. Trying to balance two systems that weren’t meant to balance might be an interesting use of time but is in the end, futile.

    The suggested use by others on this thread of proportional analysis is ridiculous, IMO. It’s also a stretch of the law. Rossi has to prove that any felons who voted did so in such a way as to pull the election to Gregiore. The only way to do that is to take sworn affadavits from the felons who illegally voted.

    That’s the bed Rossi’s made for himself.

  35. 38

    Brent spews:

    Goldy wrote, “The truth is, we had a run-of-the-mill imperfect election, with the normal assortment of randomly distributed errors.”

    Boy golly, with the numbers fully admitted by Mr. Logan himself with regards to more votes than voters, compared to that of previous elections, make a statement like this pure folly. NORMAL assortment? Did you even look at the numbers??

    It doesn’t matter what side of fence one sits on, the numbers, which have been admitted to, just don’t add up compared to previous elections. Speculation and theories can be made, but the numbers are the only concrete evidence we have to show any proof.

  36. 39

    Goldy spews:

    Brent @38,

    And what numbers, pray tell, are you comparing this election to? The “about 20″ discrepancy that Stefan touts over on (u)SP? Has he ever documented that? Is there any record of this number? No… it’s just hearsay, as is most of his “evidence.”

    You believe it it is true because you choose to.

  37. 40

    spews:

    angry @ 36
    who said anything about it being an issue? I said it wasn’t near sufficient to set aside an election. Is it an issue to discuss for improving subsequent elections? You bet. Is it an issue for deciding the last one? Not a chance.

  38. 41

    spews:

    angry @ 36
    who said anything about it being an issue? I said it wasn’t near sufficient to set aside an election. Is it an issue to discuss for improving subsequent elections? You bet. Is it an issue for deciding the last one? Not a chance.

  39. 42

    Tom Rekdal spews:

    I check into your postings fairly regularly in an effort to keep up with the debate between you and the indefatigable Mr. Sharkansky, but it is getting harder and harder to follow the most interesting threads through all the personal invective thrown around on all sides.

    Erik @ 8 puts his finger on what is still the sticking problem for me. It is not enough to show that there have been more irregularities and invalid ballots than the margin of victory. There needs to be a theory of how much is too much, and compared to what? I just don’t see that yet, as much as I would like to see Rossi as Governor.

  40. 43

    Dubyasux spews:

    Pot Calling Kettle Black Department

    I heard from a lawyer involved in the contest that several of Rossi’s counties didn’t check provisional ballot signatures at all and nearly 2,000 provisional ballots were counted in these counties without verifying the eligibility of the voters.

  41. 44

    zapporo spews:

    Carla oh Carla,

    Just hold on to Toto, click your heels and soon you’ll be back in Kansas with Aunt Mae.

    First, you are seriously delusional if you think Chrissy doesn’t realize this election is a bust.

    Second, when Curly, Larry, and Moe are in charge of elections in the largest county in the state, well, Judges tend to frown on numbers that don’t quite add up. Especially Enron-accounting style numbers.

    I bet it just burns you to think that this state is slowly but surely becoming more conservative just because of the excesses perpetrated by the party of the big donkey. Hee Haw!

  42. 45

    dj spews:

    Dubyasux @ 43

    Well, DAMNATION! That does it . . . errors is errors. Now we really have no idea who REALLY won the election. I’m gonna hafta jump on the REVOTE bandwagon. I hope somebody over at (u)SP is scrutinizing their LITTLE BINDERS. :-)

  43. 46

    dj spews:

    torridjoe @ 32
    Spot on! My analysis is only applicable to the number of ballots that we would, ideally, like to remove from the totals (but cannot remove directly). I trust that even our most empassioned conservative friends would accept that the improperly cast provisional ballots should be kept so long as the county subsequently verified that the voter was eligible.

  44. 47

    zip spews:

    Don @ 16, Goldy, chew2, dj

    Thanks for not responding to my comment (again). Maybe I should change my irregular handle to “Don lies”. You guys should give dancing lessons.

    Don, you said that the anecdotal evidence indicates that a majority of the felon votes went to Rossi. Are you saying that your anecdotal evidence indicates something here? On a “more probable than not” basis? I still object to you and the regular lefties here repeatedly throwing that claim out as if it were fact. You said:

    “If it turns out the majority of illegal felon votes went to Rossi, as the anecdotal evidence indicates…”

    Come on and admit that the anecdotal evidence does not indicate a darn thing. Just like the ASA link doesn’t indicate anything. I’d hang my hat on the ASA pattern of felon votes before I hang it on your anecdotal evidence. In my view, it’s not “evidence” after it’s been filtered by the press.

  45. 48

    Dubyasux spews:

    Zip, the anecdotal evidence I cited is newspaper reports that, as I recall, something like 7 or 8 ineligible felons said they voted for Rossi and 1 said she voted for Gregoire. That’s not enough to prove a legal case or provide a statistically valid sampling, but right now it is the only factual material we have. Everything else is conjecture. However, common sense would corroborate this trend as most felons are males and we know the male vote trends Republican. Far more significant to me is the fact several Rossi counties didn’t check signatures or verify the eligibility of provisional voters, they simply assumed these ballots were valid and counted them. It doesn’t take a rocket scientist to figure out this is the death knell for any Republican hopes riding on proportional analysis, even if they can get Judge Bridges (and the Supreme Court) to go along with proportional analysis. What the Democratic attorneys have is three times as many improperly canvassed provisionals from Rossi counties as there were in King County. Not only that, in KC Gregoire led Rossi by only 18 points, whereas in many of the Republican counties Rossi led Gregoire by twice that margin, so you are really looking at a 6-to-1 ratio of improperly canvassed provisionals in favor of Rossi. Rossi’s felon list has already imploded; this wipes out any gain Rossi might have got from the KC provisionals under proportional analysis, and puts him in negative numbers. According to the Democratic attorney I spoke to, the GOP game plan now is to keep illegal votes for Rossi from being admitted into evidence so that Judge Bridges will only have the illegal votes for Gregoire in front of him. This shows you how dishonest and cynical the Republican litigation strategy is. It proves they aren’t interested in finding out who really won the election, they’re only interest is in putting Rossi in office, even if he lost. Shame on them.

  46. 49

    zip spews:

    Seems reminiscent of the dishonest and cynical Berendt-led affidavit gathering effort in November.

    If your facts are correct, you may be right that it spells curtains for Rossi’s case. I’ll keep my fingers crossed until the true facts come out.

  47. 50

    Brent spews:

    Goldy @ 39,

    What facts do you have to show otherwise, other than harsh words?

    I read both sites. I don’t consider myself for one or the other. If you want to disprove it, DO SO! You can disparage something/someone through words, or you can blast them with reality through cold hard facts. Which do you think is more convincing?

    I may choose to believe it, but what choice is there when no other case is presented? It’s a compelling case if you look at it objectively. I’m not tied to any donkey or elephant, so lay out your case why it’s all just hearsay. Prove it!

  48. 51

    Chee spews:

    zap@49. It is a big leap between what we are told and what we actually see. So many numbers, so many errors and so many different spins and slants. All in all taken with a grain of salt, I doubt of the court(s) will ever favor a revote aka re-election. Not because there were errors, not because there were not errors in the past elecion and not because it can be proven either candidate won or lost. But because. overthrowing an election at the state level, would bring about many varibles, changes in voter turnout, changes in minds, changes in residancy, people who have moved out of state, changes in death rolls and changes in voter rolls re new voters turning 21; to name a few. The cleansing of the unwanted aka uneligble would be a factor in change also and I have no arguement there. To avoid this kind of mayhem, the law has set a high bar. If Rossi wins and I do not think he will, we will all have to live with it as we live with Duyba. Gasp for air!

  49. 52

    dj spews:

    zip @ 47
    Sorry, zip, I though I had answered your question: “How about some reason why the Washington felons would differ so much from the national pattern?”

    1. The national pattern for felons is not known. It was inferred. The authors played a “what if” game: IF (ex)felons voted just like non-felons when matched for a bunch of demographic characteristics (sex, race, education, age, etc.) , THEN what affect would allowing (ex)votes by felons have on some previous elections.

    2. Washington felons differ from national felons by at least one important demographic characteristic: the proportion of black felons. Nationally, African Americans tend to vote Democratic (11% voted for Bush, 88% Kerry in CNN exit polls). Washington has 15% black felons, U.S. average is 50% black felons.

    That is why Washington felons likely differ from the national pattern.

    “Come on and admit that the anecdotal evidence does not indicate a darn thing.”

    Anecdotal evidence is just a small collection of anecdotes–the information was collected unsystematically, so it makes a nice teaser. . . . By adding “Anecdotal” in front of the word “evidence,” we are really saying “it does not indicate a darn thing”. There. Does that confession work for you?

  50. 53

    Goldy spews:

    Brent @50,

    As usual, I am being asked to prove a negative. That’s not how logic, or the courts, work. According to statute and precedent, the presumption is that we have had a clean and fair election. It is up to the other guys to prove otherwise.

    If you believe that this election is so much less accurate than previous ones, prove it. But to base your evaluation of the 2000 election, as Stefan does, on the recollections of David Irons, the man challenging Ron Sims, without any documentary evidence to back you up, just isn’t compelling.

  51. 54

    Chee spews:

    A briefing states, Washington’s Secretary of State Sam Reed will be speaking in his home town at the Wenatchee Valley Chamber of Commerce’s Wake Up Wenatchee event, a breakfast event held from from 7:AM to 8:15 at the Red Lion Hotel, 1225 N. Wenatchee Avenue on April 26th. Reed will address the 2004 governor’s race and state election reform.

  52. 55

    Marilyn spews:

    Hi there! This is my first read of this site, and I have to say, I do enjoy it. Re the “felons” list, and the GOP assertion that it is possible to statiscally extrapolate that most of them must have voted for Gregoire, I have an alternative assertion: Most felons would have a natural affinity for Dino Rossi and probably most felons voted for Dino. The reason: Washington State is mostly vanilla, and most felons are male, and since most felons in Washington State are likely white males (sorry guys) most of the felons who voted probably voted for Dino Rossi, after all he’s the “manly” choice. Most probably wouldn’t vote for a woman, and certainly not for a former attorney general. So, most felons voted for Rossi, no doubt about it. Now, I haven’t confirmed that assetion with any further research, but then the GOP doesn’t require a very high standard anyway, and I think my “research” is equal to their standard. So..it’s clear to me – most felons are Rossi fans and voted for Rossi, which explains Rossi’s initial “lead” or “win”. Now we know. when felons vote, Republicans can “win”. What do you think? Marilyn

  53. 56

    Dubyasux spews:

    zip @ 49

    The Democrats did nothing the GOP hasn’t done themselves in the past, and it’s perfectly legal. I can understand why the Repubs are beefing about it though — it’s pretty embarrassing they didn’t cover their asses on that one.

  54. 57

    Dubyasux spews:

    Marilyn @ 55

    That’s probably where it’s going, Marilyn. Of the felons who talked to reporters, all but one said they voted for Rossi. The GOP seems to realize it’s not to their advantage to take actual testimony from the illegal voters because they’re arguing against it and have already begun to spin that “you can’t believe what a felon says” in case it happens. They want to focus on illegal votes in King County and ignore illegal votes in Rossi counties. They’ve made much trumpet of 660 provisional ballots that KC voters fed into voting machines, but you don’t hear them talking at all about Rossi counties that didn’t even check provisional signatures. They’re playing games, and the public sees through it. Rossi’s poll numbers have plummeted and three-fourths of the state now opposes a revote.

  55. 58

    ChrisN spews:

    dubyasux @57

    What is the total number of provisionals in the rossi counties that didnt get checked?

    How many felons were interviewed by the press and reported to vote for rossi? (these votes are illegal, and should be disgarded regardless of whom the vote was cast)

    How many illegal votes were in rossi counties?

    Of the 660 provisionals fed into machines, how many were legitimate? How many were counted twice (once as absentee, once as provisional)?

    In rossi counties, what is the total voter/ballot discrepancy, between all the counts?

    If you have the answers to these questions, the picture becomes clearer…… The numbers will tell the truth…. Opinions dont count.