King County Elections has now had the petitions for I-24 and I-25 for 28 days. According to their status update, they won't start on I-25 until they're done verifying I-24. At this point, they're less than halfway done checking I-24 and plan to spend the rest of the month examining tens of thousands more signatures, even though it's already a statistical sure thing to qualify. At the current rate of progress, it's impossible for them to accomplish a full check on I-25 before the Council's September 21 deadline.
The waste of time and money and disregard for the public's interest in a reasonably prompt qualification of the initiatives is obscene.
It turns out that the legal authority for using statistical sampling is more complicated than I reported earlier King County Code 1.16.100 (here, p. 47) allows the Elections office to "use any statistical sampling techniques for this canvass which have been approved by the county council." Oddly, while the council authorized the use of statistical sampling in principle, the Clerk of Council cannot find any record that the council approved any specific sampling methodology. (my e-mail exchange with the Elections office and council staff here). Former elections managers who had told me earlier that they had used sampling in the past were apparently thinking of Seattle initiatives, while the county code section applies only to county initiatives. In practice, there haven't been very many countywide initiatives that turn in enough signatures to qualify where sampling would be appropriate. It hasn't happened in years (if ever) that two initiatives were turned in at the same time.
But Ron Sims/Sherril Huff's response to the situation is disappointing. The petitions were being processed by the Clerk of Council for nearly a month before they were transmitted to Elections. Sims/Huff should have anticipated the workload and sent an urgent request to the Council to approve the same statistical sampling method that the Secretary of State uses. And they could have tested a sample upon receipt to demonstrate to the Council the certainty of qualification and the foolish waste of performing a full check. But they didn't. Instead they're spending an unnecessary several weeks and $50,000 in taxpayer funds to drag this out.
Elections spokeswoman Bobbie Egan sent me this progress update on Friday afternoon:
King County Elections completed the second week of signature verification on I-24. The last update I received was Tuesday and at that time we had processed approximately 17,161 signatures and verified 13,197.I-24 turned in nearly 82,000 signatures and therefore needs only a 67% validity rate to meet the qualification threshold of 54,732. The 17,161 sample had a 77% validity rate. No matter what you think about I-24 itself, it's clearly met the qualification threshhold and there's absolutely no good reason to continue checking its signatures. (In statistical lingo, I calculate that the proportion of valid signatures in the sample is 31 standard errors away from the null hypothesis of 67% validity, so the confidence level is essentially 100%). And if a random sample of 3,000 I-25 signatures had a similar 77% validity rate, you would have 99.997% confidence that its 74,200 signatures contained at least the target (54,732) of valid signatures. (i.e. the probability of not really having enough valid signatures, but getting a fluke random sample showing a 77% validity rate, is only 0.003% or 1/33,000).The verification process is steady and expected to reach completion toward the end of August. It is not practical to give a specific date of completion given the coordination of this work with the August Primary.
But instead of going to the Council and putting the ball in the Council's court to authorize sampling to save the taxpayers tens of thousands of dollars, Sims/Huff will spend the money to drag this thing out for as long as possible. Statistical sampling aside, they're still proceeding at an irresponsibly slow pace. It's not clear from Egan's status report exactly how many business days it took to count the 13,197 valid signatures, but I generously read it as 6 (first thing Tuesday of the 2nd week). That means they're counting at most 2,200 valid sigs a day, and if the process is "steady", it would mean they need at least another 46 days (after Tuesday, July 31) to finish both petitions at that rate. (each initiative needs 54,732 valid signatures, and they've made up the rule that they will count an extra 5% for each one, i.e. a total of 114,937 valid signatures; and they had reached 13,197 on Tuesday, July 31, so they need 101,740 more). There are only 23 business days in August and 14 in September up to and including the Council's deadline for action. So the schedule isn't plausible, even before contingencies for delays due to the primary.
The way this is being handled smells like either (1) stunning incompetence by elections administrators; or (2) a cynical game plan to run down the clock to (a) spare the Council from having adequate time to debate the measure or hear from the public, and/or (b) falsely reject the petition without giving the campaign enough time to appeal before the deadline to get it on the ballot.
Posted by Stefan Sharkansky at August 06, 2007 10:45 AM | Email ThisAsk each one of them why they buried the story. (They tend to hate that.)
Posted by: johnny on August 6, 2007 12:26 PMSounds like someone is bucking for more money for their department. Someone taking classes from the Queen?
Posted by: swatter on August 6, 2007 01:44 PM"The way this is being handled smells"
Sooner or later, the petition will go on the ballot. The hard part is done. All you are hearing now are the dying screams of the partisans.
Ken
Dallas
However, I agree with you that the council should approve random sampling techniques for both initiatives, and if a full check is necessary (either because sampling shows it will be close, or because the council doesn't approve sampling), then enough resources should be obtained to do it more quickly. And I say this even though I am inclined against both initiatives.
Posted by: Bruce on August 6, 2007 02:05 PMThe legislature seems hell-bent on wiping out the Constitutionally protected initiative process. Government cities and counties are of the same frame of mind.
Suppose, just suppose that KCE cannot finish the validation in time for the election. What then? Immediate action by the legislature to minimize the whole process because the cost to validate petition signatures is too high? I think a whole bunch of worms pop out of the ground in the favor of destroying the initiative process will occur.
Anybody else concur?
Posted by: swatter on August 6, 2007 02:33 PMHe does have an obligation to ensure that such things are taken care of in a timely manner - failure to do in such a deliberate manner isn't going to be endearing him to anybody.
Posted by: John Galt on August 6, 2007 03:25 PMConceivably, I could vote for Christine Gregoire. Likely? Not a bit.
Do you have a shred of proof that signatures obtained later have lower validity rates than those obtained earlier in any other initiative, or are you just making stuff up for the sake of argument?
Posted by: Palouse on August 6, 2007 03:41 PMAt the end, I think she did something like double the money that Ecology got every year. They at least doubled the staff.
Posted by: swatter on August 6, 2007 04:11 PMhttp://www.secstate.wa.gov/documentvault/MostFrequentlyAskedQuestionsonI917-1558.pdf
Did Fruto Boy Crispila sign either initiative?
Posted by: Richard Pope on August 6, 2007 04:19 PMIf one believes their signature may have been eggregiously rejected is there any recourse?
Posted by: daveo on August 6, 2007 04:46 PMWhy has this former King County Director of Records & Elections never proposed any statistical sampling method for approving initiative signatures? This person has had almost two decades to either recommend (as elections director) or implement (as council member) a reasonable statistical sampling method.
In any event, this matter is strictly up to the King County Council. If you don't like what the incumbents have done, you should elect someone else in their place. :)
Posted by: Richard Pope on August 6, 2007 05:15 PMIf a person signs more than once on a state petition, the first signature is counted and subsequent signatures are rejected.
But if a person signs more than once on a King County petition, all the signatures are rejected, even the first signature submitted by that person.
This can be very significant, given that some state petitions have had up to 6% duplication rate. On a state petition, that would knock out 6% of the signatures -- i.e. the second time the person signed. But on a county petition, it would knock out 12% of the signatures -- both the first signature and the second signature of each double signer.
This can CERTAINLY be managed with a statistical sampling method, and Stefan could come up with an algorithm (or even an Al Gore Rhythm) for the different rule in King County.
Posted by: Richard Pope on August 6, 2007 05:24 PMBut the percentage of duplicates will grow exponentially as the number of signatures counted increases. For example, with I-917, there were only 0.2% duplicates when 4% of signatures were counted, but 5.45% when 100% of signatures were counted.
Considerably that duplications means two signature rejections (including the first signature) and not just the second signature rejected (with the first one counted under the state system), this could very easily result in I-25 failing to qualify.
In any event, it should be very close in qualifying, and statistical sampling methods might very well have required a 100% count in the first place for I-25.
On the other hand, I-24 might very well have qualified based on statistical sampling methods, since it turned in 8,000 more signatures than I-25.
Posted by: Richard Pope on August 6, 2007 05:33 PMYou are overestimating the likely number of duplicates. First they do not "grow exponentially", the number of duplicates in a sample is expected to increase quadratically = proportional to the square of the sample size (think about the number of unique pairs of signatures in a sample of N signatures, which is N(N-1)/2 ). The percentage of duplicates that is reported in the Secretary of State's document is not a true rate, but was calculated by simply extrapolating from the number of duplicates in the sample to the size of the entire batch of submitted signatures.
Statewide initiatives submit approximately 4 times as many signatures as I-24 and I-25, so you might expect 16 times as many duplicates in the submission, or a percentage of duplicates/signatures that is about 4 times as high as in a King County initiative. (Yes, this is a simplifying assumption to illustrate a point)
Posted by: Stefan Sharkansky on August 6, 2007 07:53 PM"Quadratically" is the correct word. I took "exponential" from Secretary of State Sam Reed's submission on I-917, that I linked above.
The percentage of duplicates will increase in direct relation with the number of signatures, while the number of duplicates will increase in quadratic (square) relationship with the number of signatures.
As for the percentage of duplicates, the smaller number of signatures required for King County initiatives (about 1/4 of the number required for state initiatives) must be balanced against the fact that the target population (King County voters) is only about 30% of the number of state voters.
Posted by: Richard Pope on August 6, 2007 08:29 PMNo. While state petitions have 4 times as many signatures, those signatures are drawn from 4 times the population. Therefore, assuming similar signature-gathering practices, you would expect the same percentage of duplicates in county and state petitions, since each has the same percentage of its population.
Of course, signature-gathering practices vary with every initiative, so this calculation is just a starting point.
Posted by: Bruce on August 6, 2007 08:37 PMOf course, we shouldn't have to guess any of this. A good random sampling would eliminate these factors, and that's what the elections office should do. All I'm saying now is your numbers are unfounded.
Posted by: Bruce on August 6, 2007 08:44 PMRon Sims gets away with his BS because YOU let him!
Posted by: grandgadfly on August 6, 2007 10:24 PMThis is about the "ins" and the "outs".
Posted by: John Bailo on August 7, 2007 08:12 AMHere's a newsflash for you. Litigation costs money, alot of it. Perhaps you'd like to fund this lawsuit, eh? [crickets]
Posted by: Palouse on August 7, 2007 08:14 AMOne of the people involved with the initiative could file. I notice how quick you are to slink into the excuse corner Palouse. Loosers find excuses why not to do something, winners do something. Thank god our founding fathers were men of action rather than men of excuses like you Palouse. Pathetic snivelers.
Posted by: grandgadfly on August 7, 2007 11:16 AMAnd what exactly have YOU done to stop Ron Sims, gadfly (man of action)? Nothing? Then stop bitching about what others are or are not doing. Those that filed that initiative have done more than you will ever do to affect public policy.
Posted by: Palouse on August 7, 2007 12:30 PMPerhaps, Palouse, the gadfly thinks the lawsuit ends once you file and the other side will automatically rollover; in reality, the fun and the payments are just beginning.
Posted by: swatter on August 7, 2007 01:50 PMI doubt gadfly is Richard Pope. I think Richard is smart enough to know the difference between the word "loose" and "lose".
Posted by: Palouse on August 7, 2007 02:51 PMWhile state petitions have 4 times as many signatures, those signatures are drawn from 4 times the population. Therefore, assuming similar signature-gathering practices, you would expect the same percentage of duplicates in county and state petitions, since each has the same percentage of its population.
That's a reasonable point. When I wrote that the statewide initiative would have 16 times as many duplicates as the county that was based on the assumption that in both cases there's a sample of the identical size with an identical number of duplicates.
There are enough differences in the way that signatures for the two kinds of initiatives are gathered that you can't accurately predict the duplicate rates of one based on the other. (e.g. statewide initiatives have almost twice as much time to collect signatures. That by itself would likely increase duplicates in state petitions).
Richard: A 1970s State Supreme Court decision (88 Wn.2d 247, SUDDUTH v. CHAPMAN) overturned as unconstitutional a state law that rejected all signatures from a voter who signed a petition multiple times. I assume that would invalidate the clause in King County Code that 1.16.100 that does the same?
Posted by: Stefan Sharkansky on August 7, 2007 03:37 PMI sincerely doubt it.
Posted by: GS on August 7, 2007 04:12 PMSection 1.16.100 of the King County Code was last amended in 1975, prior to the Sudduth decision.
The first proposed ordinance I will introduce in January will be to amend KCC 1.16.100 to bring it into line with RCW 29A.72.230. Require statistical sampling to be used, in accordance with the rules adopted by the Secretary of State. And make the first signature of duplicate signers be counted.
Posted by: Richard Pope on August 7, 2007 08:34 PM