After a brief email exchange over the weekend, I have invited Brian Hitchcock, a Kimberly Ellis supporter, to offer his argument on behalf of Ellis for CDP chair. This is offered without commentary and it’s a letter from the Ellis campaign and it’s a cut-and-pasted Word file so I hope all the charts hold up:
To: CDP Credentials Committee From: Brian Hitchcock, DSCC Delegate (AD66) Re: Chair Election Challenge – Appeal
Dear Credentials Committee members,
I am writing to you to express my concerns about the conduct of the 2017 Chair Election Challenge Review, and the Decision rendered by the Challenge Review Commission (CRC).
I shall not focus on questions of the CRC’s bias (though I cannot avoid mentioning some examples of it in their work product.) And I will not address the question of its jurisdiction. I believe that others (specifically the challenger herself, in the Appeal, as well as Greg Diamond, in his Testimony and elsewhere) have addressed both these issues better than I could.
Nor shall I address reports of specific irregularities in the conduct of the election, as I was not in a position to witness the irregularities. I will say only that I believe Greg Diamond’s eyewitness testimony is probative in that regard, and disturbing.
Instead, I shall confine myself to discussing the errors, omissions and discrepancies in the 2017 Chair Election Challenge Review, which I have observed and documented, including irregularities in the methodology, conduct and reporting of the review.
These discoveries cause me grave concerns, not only as a student of mathematics and logic, a longtime professional in the field of data processing and information technology, a technical editor and an auto-didact in linguistics, but also as a delegate who has an interest in a fair, objective, accurate and thorough review that would help to restore in me some measure of faith or trust in the California Democratic Party.
The problems that I observed during the conduct of the Review fall into these categories
- Deficiencies in form design (all phases*)
- Errors of data transcription (phase one)
- Errors of filtering, categorizing and tallying (from phase one to phase two)
- Errors of arithmetic and logic (all phases)
- Selective disclosure of key data (all phases)
- Discrepancies Not Investigated
- Conjecture and opinion presented as findings of fact (notably in the Decision)
- Maintaining an Illusion of Fairness
I shall discuss each of these in turn.
* “Phase one” is how I refer to the CRC “preliminary review” that resulted in production and publishing of Ballot Review Forms. “Phase two” refers to the testimony solicitation/ collection/ review stage that ended with the July 14th “secondary review” meeting and the creation and publication of Testimony Review Forms. “Phase three” refers to the final period after July 14th, culminating in the July 22 “final” review meeting and the preparation and publication of the Decision.
The Chair Election Challenge Review failed, not so much due to bias, but primarily due to the CRC members’ lack of skill in the principles and rigors of research, data entry and data analysis; and secondarily due to a single-minded determination to complete the review in a particular timeframe, undeterred by voluminous and continuing evidence of errors in their work product which strongly suggests serious flaws in their process.
We are all amateurs and volunteers here. And no amount of word mongering on my part or lexical, mathematical or juridical legerdemain by the CRC will make it otherwise.
But, in short… The job was botched.
Thank you for taking the time to read this Testimony, and the accompanying attachments. I am sure you will learn something. I hope you will end up convinced, as I am, that the review process was hopelessly flawed and that its conclusions are therefore unreliable; and that significant, valid, ongoing concerns exist which the CRC failed to address; and that you will rule that its “Findings and Orders” cannot stand. Sincerely,
Elected Assembly District Representative (AD66) Torrance, CA firstname.lastname@example.org
1. Deficiencies in Form Design
The problems began with forms that were in some ways difficult to fill in properly and to interpret unambiguously. Since most of the “comments” to the CRC (which I attach as part of my testimony) refer to errors on these forms, please take the time here to familiarize yourself with the forms and how they were used. (If you have already used these forms, you may wish to skim over this section.)
BALLOT REVIEW FORMS
The Ballot Review Form (referred to henceforth herein as the BaRF) was designed to represent various items of information about each ballot.
There are several opportunities for misunderstanding, due to ambiguous or misleading wording on the form itself, which led to many of the reported errors. We do not know how many errors remain unfound.
Ballot Review Form — Proxy data (top section)
The Proxy section was a bit complicated. Filling it out properly depended on knowing how to decode “DSCC numbers” (badge numbers). That is, people filling out the form had to recognize by the badge number whether the delegate was a county committee member or an elected “ADEM” delegate (as opposed to several other kinds of delegate), to know whether they should check if the proxy lives (and is registered as a Democrat) in a particular Assembly District or county, and which AD or county to check. In some cases proxies were coded as invalid and the form was later corrected to reflect a valid proxy, showing there was misunderstanding as to how and when to apply the validation criteria. Also, the Proxy section has a checkbox pair with a confusing negatively worded caption “Proxy Not Already on DSCC” (for which a YES answer meant a person is NOT on the DSCC and a NO answer meant that the person IS on the DSCC.) Some errors occurred on this item.
Ballot Review Form — Ballot and Signatures data (middle section)
In the middle section, the choices for the “Ballot Cast” item are Yes No (With Blank) No (No Blank)
In the “Chair” item, there is also a choice Blank
In several cases, it appears there was confusion as to how to fill these out and what the combined options might mean, because some BaRFs were then “fixed” by a second quality-control reviewer.
Should a BaRF coded as Yes for Ballot Cast but Blank for “Chair” mean something different from one coded as No (With Blank) but with no box checked in the “Chair” item?
Should a BaRF coded as No (No Blank) for Ballot Cast but Blank for “Chair” mean something different from one coded as No (No Blank) but with no box checked in the “Chair” item?
What does No (No Blank) even mean? Does it mean that a blank ballot did not exist?
My query to the CRC about how these items were to be properly filled out, and what the various combinations might mean, went unanswered.
Thus there seems to be no way to know, based on what the CRC reported, how many people cast ballots without voting for Chair; that is, how many abstained from voting for Chair. Whether this obfuscation was intentional, I do not know.
Preliminary Determination Section
Ballot Review Form — Preliminary Determination data (bottom of form)
This is where the overall determination as to whether the ballot needs further review is indicated. This is the most critical point in the entire three-phase process, as if a mistake is made here that indicates a ballot is valid, when it isn’t, the ballot might never be looked at again. In my attached comments, I pointed out to the CRC some of that very type of error.
There are three possible choices for a Preliminary Determination.
VALID VOTE/ INVALID VOTE COMMISSION NO VOTE REVIEW
There is also a “Review by” checkoff to indicate which member of the CRC reviewed the ballot.
Combining the two distinct situations “valid vote” and “no vote” into one checkbox led to difficulties in reporting (or should I say ease of obscuring) how many ballots were voted and how many were not See Selective Disclosure of Key Data, below.
But this sleight-of-hand did make it easier to set aside all the ballots determined to be VALID in the preliminary pass (which the CRC disingenuously claims was ONLY preliminary), lumped in with the unremarkable unused ballots, never to be looked at again—unless someone complained.
TESTIMONY REVIEW FORMS
The Testimony Review Form (referred to henceforth herein as the TeRF) essentially duplicates the Ballot Review Form, with the addition of a small portion at the top to indicate “CRC Action items” and to very briefly recap testimony received in regard to the ballot.
Testimony Review Form — Testimony section (top of form)
This area also has a space to record the “Previous Prelim Determinations”. However, this space was not used to show the overall preliminary determination for the ballot (VALID / NO VOTE, INVALID, or COMMISSION REVIEW); it was used to describe specific problems found in the preliminary review.
The middle section of the form is a slightly reduced image of a BaRF.
Testimony Review Form — mini BaRF (middle of form)
Any changes made relative to the original markings of the BaRF, based on review of the testimony, are supposed to be circled.
Testimony Review Form — Notes area and Determination (bottom of form)
This is where notes are added to indicate new information (whether exculpatory, condemnatory or merely explanatory) or any new decisions . And, if the overall determination changes, the checkbox for the “new” determination should be circled.
However, as most reviewers chose not to indicate the prior preliminary determination in the top section of the TeRF, one cannot tell from the TeRF alone what the actual change was, only the result of the change (except when the only change was to add a note, or where the prior state of a checkbox or other item is explicitly stated in the top portion).
So, for example, if the preliminary determination were completely reversed from VALID VOTE to INVALID VOTE, the marking on the TeRF would be the same as if an existing objection from the preliminary determination was essentially confirmed, by changing COMMISSION REVIEW to INVALID.
As we shall see later, TeRFs were not even created for ballots where testimony review resulted in a confirmation (upholding) of the preliminary information & determination as found on the BaRF. Note that I refer to such cases as Testimony Review Determinations (or TeRDs for short).
Final Review “Forms”
Oops, there weren’t any. I asked the CRC when forms would be provided. It seems they will not be.
The results of the July 22 deliberations were reported in the form of a long narrative included in the Decision, with varying degrees of disclosure of details.
Some decisions applied to an entire “category” (issue type) and others to individual ballots. Although totals were given for each of the four categories of “signature match issues”, and 26 other categories, including missing signatures and a multitude of proxy problems. For most of these categorical decisions, there was no indication of whether the final determinations resulted in reversals of preliminary determinations, or confirmations of preliminary determinations (or a mixture. As I noted earlier, not all reviewers had applied the same preliminary determination for the same problem.)
As noted above in the discussion of BALLOT REVIEW FORMS, it is not known whether consistent guidelines were applied as to just what circumstances called for flagging a BaRF as needing Commission Review, and which circumstances called for flagging it as Invalid. However, the severity of any one issue, or especially of various combinations of issues, was not known in the first phase; that was one main purpose of the third phase.
Nor were the individual voters identified for most of these categories, even when there was only one voter in the category!
One has to dig to figure that out. As it turns out, there were ten outright reversals of VALID to INVALID (seven for Ellis, three for Bauman). This might have been evident if there had been any Final Review forms. I have yet to determine how many were reversed from INVALID to VALID. Certainly more than ten.
2. Errors of data transcription
First Pass – Ballot Review Forms
The first pass of the review was conducted starting some time in June (the start date of the review, if ever announced, is not available now) and ending on July 3rd. This resulted in creation of more than 3300 BaRFs.
The CRC released these on July 3rd, amid a long holiday weekend. I was made aware of that release on the evening of Wednesday July 5th. The CRC set an extremely short window for comment, to end on Monday July 10th. Therefore, I had only two business days and a weekend in which to gather a team of researchers to examine more that 3300 BaRFs. My team (consisting of fellow delegates Margie Hoyt, Lisa Kato, Diana Parmeter, Lucio Reveles, Naida Tushnet, Yosh Yamanaka, with additional help from Rashelle Zelaznik, Chris Robson and Bill Roberson) were able to examine 13 of the 40 numbered BaRF pdfs that the CRC (files 1, 2, 9, 10, 12, 15, 17, 19 and 24, containing approximately 1000 BaRFs, as well as the individually named BaRF files published along with them.
Based on this brief, partial examination, we found and I reported 12 errors on the BaRFs, in “timely” complaints submitted to the CRC. See Attachments 1 through 12. Keep in mind that my team had no access to any of the source data (that is, the ballots, the proxy forms, the ballot sign-in sheets, or any other ancillary material.) We could see only the derivative “metadata”, that is, the BaRFs that the CRC had filled out. So the only types of errors we could detect were those that presented obvious omissions, ambiguities or inconsistencies in the way the form was filled out. We had no way of knowing how many additional BaRFs might have been incorrect representation of the data, but which did not show omissions, ambiguities or inconsistencies.
For example, we could see if the reviewer had erroneously failed to fill in a required checkbox to indicate whether the ballot had been cast, or whether a signature had been entered, or whether a proxy was actually a registered voter from the right district, but we could not detect whether a box might have been filled in erroneously for an item that had not been checked, as long as all required entries were made on the form. That is, as long as the BaRF was completed, and in a way that looked consistent, there was no way to tell if it had been completed erroneously.
Or to cite another type of example: we encountered more than one BaRF where a voter had reportedly used the wrong ballot, and yet the BaRF for the ballot that should have been used showed the vote, while the BaRF for the ballot that was actually used showed no vote. In other words, data for one ballot was recorded on the BaRF for another. This happened most notably in the “SoS” case.
Anyway, as the brief comment period ended on the 10th, we had searched only 13 of the 40 posted pdf files of alphabetically-sorted BaRFs. Nevertheless, I continued to search, certain there were more errors to be found. I continued to find and report errors, and continued to complain that the allotted comment period was insufficient. This allowed me time to collate my team’s existing results and to examine a couple more files on my own, for a total of 15 files (about 40% of the BaRFs).
During that week, I reported nine new errors and submitted four additional comments and queries, all of which were labeled as “late”. See Attachments 13 through 25.
Eventually, the CRC relented and “extended” the comment period by another week, to the 17th. However, they still labeled all submissions received after the 10th as being “late”.
Meanwhile, another larger team, unbeknownst to me, had been likewise examining the BaRF files, and had completed their examination. This team included, among others, delegates Cari Templeton, Katrina Bergstrom, and Greg Diamond.
Katrina Bergstrom submitted a single statement, organized by error type, that raised various concerns, mentioning 194 specific instances; approximately 60 were specific errors on the BaRFs; others were comments re consistency of the determinations, and comments as to whether certain types of errors should or should not have invalidated a ballot. Katrina also identified some inconsistencies, wherein apparently identical situations resulted in a different Preliminary Determination. In addition, Katrina noted some of the same ambiguities in the forms as I detail above in Deficiencies in Form Design. See Attachment 26.
Greg Diamond submitted, on July 10, a single comment listing various problems and errors affecting 14 distinct ballots. See Attachment 27.
Greg Diamond then submitted another comment, and a follow-up, in regard to an unspecified number of ballots flagged as INVALID for non-payment of the convention fee, which Greg argued based on the Bylaws should not have been flagged as INVALID (note that eventually, the CRC validated 7 such ballots. So we could say Greg had complained about 7 ballots.) See Attachments 28 and 29.
So between the three of us, we made 236 specific comments errors or anomalies evident from the BaRFs, including more than 100 specifically about errors on BaRFs.
There was, of course, overlap between the errors the two teams found. But interestingly, there were some errors that my small team found in our partial sample which the other team had not found in examining the complete set of BaRFs.
Thus, the more eyes, the more errors detected. Given more time and more eyes, and access to the actual source data, who knows how many more errors might be detected. Sadly, if an audit is not ordered, we will never know how many “substantial issues” are lurking in the “Consent” pile.
3. Errors of filtering, categorizing and tallying
The CRC stated that the purpose of the BaRFs was to “identify votes that may be potentially invalid or required additional review by the CRC…”
The BaRFs failed to do this. Some ballots that should have been tagged for further review were not; other ballots that should not have been so tagged, were. And some were indicated on the BaRF as both Valid and Invalid, or Valid and [needing] Commission Review. So they do not serve as a valid means of filtering out the “good” ballots from the “questionable” ones meriting further investigation. This is the crux of the issue, as demonstrated in Errors of data transcription, above.
Nevertheless, despite having been notified of dozens of errors on the BaRFs, the CRC proceeded to the second and third phases, as if the BaRFs provided a valid filter. They stated in the Decision that the BaRFs “captured multiple pieces of information” when in fact it was the CRC members themselves who captured, or failed to capture, the relevant, necessary information, and transferred it with varying degrees of success to the BaRFs, resulting in an error rate which I estimate to be greater than 2% (I found 22 errors out of 1000 BaRFs). This could balloon if we count the uninvestigated “stealth” delegates (mentioned in item 7 below).
There is a complicated error I did not discover and analyze until the week of the “final review” meeting July 22.
According to one BaRF, the Secretary of State (Alejandro Padilla, DSCC# D003000) did not cast a ballot. See attachment 30.
But according to another BaRF, delegate (Alex Padilla, DSCC# D354004) cast a ballot but did not pay his registration fee. See attachment 31.
However, according to sworn testimony apparently submitted by the Secretary of State, who calls himself Alex Padilla, he did cast his own ballot but did not pay a registration fee (and didn’t seem to think he needed to.) See attachment 32.
This testimony was alluded to in the Testimony Review Form prepared for DSCC #D354004. See attachment 33.
There was no testimony from the other Alex Padilla as to whether he had cast a ballot, or paid his fee—perhaps none was solicited.
I reported this error on July 23rd, and the CRC noted their error in their Decision published on July 26th. Mr. Wagaman also explained (on the CADEM 2017 Delegates Facebook page) that this had been entirely the CRC’s error, not a case of mistake on the part of the election volunteers either in issuing the wrong ballot to the Secretary of State or in issuing the wrong “credential” to his staff, as one might have inferred from the usage of the wrong BaRF.
So this is the nature of admitted errors by the CRC. And the mix-up between Alejandro Padilla and Alex Padilla was apparently not discovered by the CRC during three rounds of review, as it was acknowledged only after I pointed it out to them, after their “final” review.
4. Errors of arithmetic and logic
197 HAD “SIGNATURE” COMPARISON ISSUES ONLY, 140 HAD “OTHER” ISSUES ONLY, 18 HAD BOTH KINDS
The CRC claims (in the Decision, bottom of page 9) that “Of these, 197 involved only signature comparison issues.…140 involved only issues not related to signature comparison….18 involved both signature comparison and other issues”
However, this does not agree with the number of voters listed in the “Voter List-V2” file that the CRC has made available via a link on its web page (incorporated herein by mention.) Proposed Consent and CRC Review Issues In that file, footnotes both in the Signature Issues section and in the Other Issues sections, state that names with both signature-match and other types of issues are indicated by a plus-sign “+” in front of the name. On both sections, the same 19 names are flagged with “+” . Not 18.
The total names listed in on the Signature Issues section, which should be 215 (that is, 197 plus 18 as reported by the CRC) is actually 218. Three too many. There are 199 unflagged names (that is, names with signature-match issues only). Not 197. Two too many. an error of 1%.
The total names shown on the Other Issues section, which should be 158 (that is, 140 plus 18 as reported by the CRC) is actually 156. Two short. There are 137 unflagged names. Not 140. Short by three. An error of 2.18% (the original election margin was 2.115%.) And yet we are expected to believe the CRC’s vote tallies. And their decisions based on BaRFs with a much higher error rate. There were other logic errors, but most of them were faulty syllogisms and other fallacies used to advance opinions and apologetics in the guise of conclusions, so I discuss them under Presenting conjectures and opinions as “findings”, below.
5. Selective Disclosure of Key Data
How Many Ballots Were Cast Where the Voter Abstained in the Chair Race?
As mentioned in Errors of data transcription, above, there is no way to know, based on what the CRC reported, how many people cast ballots without voting for Chair; that is, how many abstained from voting for Chair. Whether this obfuscation is intentional, I do not know.
Valid Votes Combined with Non Votes
As mentioned in Errors of data transcription, above, the BaRFs conflate “valid votes” with non-votes. Thus there is no way to know (other than by re-examining all the BaRFs) how many people had ballots but did not vote. people voted. Throughout the entire three phases of the review, and even in the Decision itself, the CRC continued to lump “valid votes” with non votes in their reported totals. The CRC never actually reported how many delegates did not vote. Whether this obfuscation is intentional, I do not know.
We are told, for example, on page 9 of the Decision that the CRC had “determined that 2,976 votes had no substantial issues…” and immediately after that, we see that this included
1) Voters identified preliminarily by the CRC as valid or not voting where no testimony had been submitted (2873 voters).
In addition to lumping apples with oranges, this subtly implies that a lack of testimony somehow confirms the validity of these BaRFs. (see item 5, above and item 8, below. This conveniently ignores the fact that testimony was not solicited for any of the 2,976 supposedly valid (and/or non-voted) ballots, so the lack of testimony on 2873 of them means nothing.
Next, we find that there were 17 other people who did not vote for chair in the election. Again, these are a mixed bag, either they didn’t vote at all, or they voted but abstained from voting for Chair.
2) Remaining voters where no vote was cast in the Chair’s contest, either because the voter did not vote in the Statewide Officer election or voted in the election but made no select [sic] in the Chair contest. (17 voters)
So that means these 17 people who did not vote for Chair were NOT initially identified as valid or as not voting. Either that, or testimony was required to determine that they did not vote. So that means that on the first pass there were 17 ballots where the CRC couldn’t determine that they HAD NOT VOTED. This is pretty hard to believe. Could you tell if a ballot was filled out or not? Apparently, the CRC could not in these 17 cases. Which cases are they? Wouldn’t you like to know? The CRC hasn’t chosen to disclose them. Whether this obfuscation is intentional, I do not know. The only way to really be sure would be to re-tally all the BaRFs.
Testimony Review – Determinations vs Forms
As mentioned in item 1 , the CRC failed to report on 31 of the 135 “Testimony review” ballots, making TeRFs only for the 104 where they reversed a prior determination on 7/14, not those where they decided NOT to reverse. (Eventually, the CRC noted their ostensible oversight, and the “confusion” it caused, but still failed to report on the 31 cases, literally failing to reveal which delegates they were. So the TeRDs remain covered up. But they still stink.
Reporting What Happened on July 22
The CRC claimed that on July 22 they “invalidated” 47 votes (25 for Bauman and 22 for Ellis). In fact, they only reversed ten prior determinations on that day. That is, they changed ten ballots from VALID to INVALID (7 against Ellis and 3 against Bauman.) They upheld the previous determinations, or confirmed the previous suspicions for the other 35 (25 of which were already coded as INVALID, and the other 10 as Commission Review).
Many ballots were redeemed (converted from questionable to valid, or even from invalid to valid). It’s hard to tell how many, because the prior (“preliminary”) determinations are reported for only some of the 355 ballots that were considered on July 22.
I do not know whether this misrepresentation and obfuscation was intentional.
6. Discrepancies Not Investigated
There is no way to know, based on the figures the CRC reported, how many delegates voted who were not qualified according to the March 1st final delegate roster. The CRC mentioned a few of them, but the only way to find them all is do this is to create a complete list based on the entire 3,301 BaRFs and to compare it with the March 1st roll of delegates.
I did so, and found 60 delegates who reportedly had ballots (i.e. they had BaRFs) but were not on the March 1st roll.
One of them, Zac Goldstein, was reported on the BaRF as an “invalid voter” (though no reason was given). The BaRF says his ballot was “voided”.
Four such delegates, whose names did not appear on the roll even when they showed up to register, were nevertheless allowed to register and vote; the CRC later declared (in a note on each TeRF) that the “Delegate info [was] properly submitted prior to 2/7 deadline”. So this “info” mysteriously disappeared sometime before March 1st, and miraculously reappeared after the election but in time for the CRC to declare these delegates valid. If that is what is claimed, the CRC seems to be attributing dereliction, if not misfeasance, to then-Secretary Daraka Larimore-Hall.
Five of these delegates, according to their BaRF, did not vote.
This leaves 53 people who were not on the March 1 roll, who did vote, but for whom there was apparently no effort made to determine whether their info was submitted before Feb 7th.
If these delegates’ information was not submitted prior to February 7th, (whether or not they were on some interim roll published after March 1st but prior to May 20th), the delegates were not qualified, as they were in violation of the Bylaws
“…The roll of eligible voters at this meeting of This Committee shall consist of those members who have qualified as of February 7, or next business day if date falls on state holiday or weekend, of the year of that meeting. …
(Article IV Section 1 (a).
In the case of an ADEM delegate, there is a penalty specified for failing to timely submit delegate information: “…withholding of that Convener’s credential…” (Article VI section 1(j)(3))
However there seems to be no penalty imposed on a County Committee Chair (such as, for instance, Eric Bauman), or an elected official (such as, for instance, Chris Parker) when they fail to notify the Chair of the selection or appointment of a delegate prior to February 7th. It seems that in those cases, only the delegate would be penalized for the county chair or elected official’s lapse.
If, on the other hand, the info was submitted prior to February 7th, we have the same situation described above for the four cases that the CRC detected; namely, that the information seems to have disappeared for up to 80 days (from March 1st to the CDP Officers election on May 20th), while presumably in the custody of Daraka Larimore-Hall, then miraculously reappeared in time for the election. (Or perhaps it never re-appeared, and thus cannot be confirmed, but the CRC did not investigate, because those delegates’ names were on the roll.) Either way, it’s odd, don’t you think? And I do not see any other possibility—do you?
There were eleven people who were on the March 1st list but had no ballot. One might presume that these people were purged from the delegate roll for good reason. However, the list of these people, and the reasons for their being purged from the roll was not disclosed, and apparently was not investigated. The only way to discover this is to create a list of delegate names for whom there are BaRFs, and painstakingly compare it with the March 1st list of qualified delegates. Which I did.
These are the delegates who were purged from the voter roll, seemingly without a trace. Maybe they died. Maybe they moved out of the state. Maybe they quit the party. Maybe their credentials were successfully challenged. Do you know what happened to them?
- D441018 Jeffrey Adair
- D441009 Shikha Hamilton
- D457008 Enrique Hernandez
- D441020 Marian Mann
- D401006 Elizabeth Nichols
- D441015 Mitchell Oster
- D419162 John Perez
- D330001 Robert Rivas
- D404003 Jenita Rodriguez
- D441019 Joe Ross
Delegate with No Ballot Who Appears on Consent List
Yes, there is one. I asked the CRC what they were “consenting” to, if this guy had no ballot. They did not reply. If you’re really curious, I’ll tell you his name. You won’t find my comment about this on the CRC web page, because the CRC is no longer acknowledging or archiving comments, since the 22nd. I hope that my comments about the errors I found and documented AFTER the “final” review have been forwarded to the Credentials Committee; I was assured that they would be. Just in case, I include the text of those email messages as attachments xx through xx It’s hard to have faith in the process, when the CRC counts as a “voter” someone who had no ballot.
For further details on the Stealth and AWOL delegates, and some other oddities, see attachment xx.
7. Presenting conjectures and opinions as “findings”
The Decision of the Compliance Review Commission (CRC) Relating to a Challenge to the CDP Officer Election for Chair Position at May 19-21, 2017 State Convention (which I shall refer to hereafter as “The Decision”), contains several dozen claims throughout the document, including a list of 36 statements near the end which are headed “Findings and Order”.
Many of the statements in the Decision seem not to be statements of fact, but rather to be opinions and conjectures. That is to say, they give the impression of rhetoric designed to frame the discussion and quell doubts, without being substantiated by facts (at least, not by the facts stated in the Decision.) And some seem designed to foster confidence in the fairness and accuracy of the CRC, without demonstrating such fairness or accuracy.
Many of the statements draw conclusions based on broad logical leaps (or fallacies), rely on unstated assumptions, or invoke facts not in evidence.
And several seem extraneous in that they are not responsive to the actual Challenge, but seem to respond to allegations apparently heard elsewhere, perhaps in Facebook chat, if at all.
I offer several examples below.
“NO EVIDENCE OF TAMPERING” The CRC claims (in the Decision, page 7) that there was
This involves two logical fallacies:
- Narrowly defining tampering as “a vote contrary to the voter’s intent”
- Assuming that any and all ballot tampering would be reported by the delegate.
What in fact was found was that there WAS evidence of a kind of tampering, which I choose to define not as altering a ballot, but as misrepresenting critical data about the ballot (by the CRC), whether by design or mistake, such that the BaRF did not accurately represent the truth of the matter. Once again, here, the CRC has admitted the fallibility of their BaRFs, which is a recurring theme throughout the conduct of the Challenge Review. QED.
* I could not find a legal definition for “ballot tampering”. The words “tamper” or “tampering” are not mentioned at all in a fairly comprehensive treatise on prosecuting election crimes put out by the Department of Justice. https://www.justice.gov/sites/default/files/criminal/legacy/2013/09/30/electbook-0507.pdf
So “ballot tampering” is apparently a layman’s term, and can mean whatever the CRC chooses it to mean. Or whatever I choose it to mean. It sounds impressive to say there was “no evidence of ballot tampering”. But as the word “tamper” wasn’t mentioned in the Challenge either, this “no evidence of tampering” claim takes on the nature of a “straw man argument”. That is, it attempts to refute something not claimed by the claimant. That is, more rhetoric / apologetics disguised as a finding of fact. See Presenting conjectures and opinions as “findings”, below .
“The Recount Confirmed…
“The recount confirmed one additional vote for Albert, along with one additional vote for Ellis and one fewer vote for Bauman. The resulting updated count of all ballots casts was: Bauman 1,492, Ellis 1,432, and Albert 8.”
The recount was occasioned by the discovery of an eighth vote for Lenore Albert. The recount, however, differed from the May 20th count not only in its count for Lenore Albert but also in its count for Eric Bauman and for Kimberly Ellis. So while the recount may have “confirmed” one thing that they hoped it would confirm, namely that Lenore Albert had received 8 votes, it also contradicted the May 20th count in regard to both the other candidates. There is thus no reason to believe the July 18th recount to be any more accurate than the original count conducted on the night of May 20th. The discrepancies cast doubt on the July 18th recount, and in my view constitute grounds for ordering a third count—so as to either verify that the May 20th count is correct with regard to Ellis and Bauman, or that the July 18th recount is correct with regard to Ellis and Bauman, or neither.
Main body of the Decision
|Page||Text||Fallacy or Other Problem||My Comments|
|6||The first phase of the review found: No Evidence of Dumping “A ballot, cast or uncast, was found for every single potential voter.”||Not responsive (straw man argument) Also not quite true.||One person listed on the CRC’s “consent” list had no BaRF so apparently had no ballot. In addition, 10 delegates who were on the roll as of March 1 had no ballots.|
|6||No evidence of systematic bias in the election or review process – “Of votes preliminarily flagged for additional review or [as] potentially being invalid, slightly less than 55% were cast for Bauman with the remainder cast for Ellis.”||word usage – generalization||evidence adduced has nothing to do with the “election process”|
|statistical / logical fallacy||Number of errors does not necessarily indicate any bias or lack of it. But the fact that the CRC found nearly as many Ellis ballots to challenge, as the Bauman ballots Ellis initially challenged, might indicate bias. In any case, it certainly does not disprove bias!|
|7||No evidence of ballot tampering “Only one piece of testimony was submitted saying the BRF reflected a vote contrary to the voter’s intent…this was later confirmed to be an error on the BRF…. “||disproving one problem with another||They are making my case, that the BaRFs have serious errors and are therefore unreliable. But apparently only when someone complains.|
|9||determined that 2,976 votes had no substantial issues…”.||word usage||These were “ballots”, not “votes”. The 2,976 figure includes unvoted ballots. (How many? Hard to say. See Valid Votes Combined with Non Votes, above)|
|unsupported claim (cannot be substantiated due to inaccuracies noted in Errors of data transcription)||What was determined, after loads of BaRF errors were found and reported, is that the first pass failed to properly identify ballots that had issues that merited further review, which was the stated purpose of the first phase. Determining whether any issue was “substantial” was not the purpose of the first phase.|
“Findings and Order”
|Page||Number||Text||Type of Problem||My Comments|
|21||4||…voters whose ballots were invalidated could potentially be validated….by the submission…of additional testimony….but under no scenario would such validation change the outcome…”||misdirection||Ignores whether votes that were [erroneously] validated could be invalidated by further testimony. There are thousands possibly enough to “change the outcome”.|
|21||14 f.||Documents related to the appointment/election of members to the DSCC….were limited to those necessary to validate the eligibility of those members who did not appear on the pre-printed rosters at convention….Challenger faille to…challenge the eligibility of any delegate…||misdirection||Fails to address the question of when, how and why 60 delegates were added to the roster after the Feb 7 deadline in defiance of the Bylaws. Secretary has a responsibility to not put such delegates on roster.|
|22||17||The CRC finds that the Blot Review Forms and Testimony Review Forms…are responsive to the Challenger’s request for information…the official list of voting delegates||Overstatement. The BaRFs are a haystack, not a list.||The BaRFs and TeRFs are not in any sense substitutes for “the official voting list of delegates” . As sources for the other types of information enumerated, the BaRFs might be a useful substitute, IF THEY WERE ACCURATE, which they are demonstrably not.|
8. Maintaining an Illusion of Fairness
As noted in item 5, above, the CRC invalidated only ten ballots on July 22nd. I do not know if it was their intent to limit invalidations to only ten. I do know, however, that every member of the CRC knew on that day exactly how many ballots would have to be invalidated to make a difference in the outcome of the election, since they had just done a recount earlier in the week.
And they might or might not have known how many had already been coded as INVALID, although they had published all the BaRFs and published 104 of the resultant TeRDs as TeRFs.
And on 7/22 they chose, through long and careful deliberation, to reverse ten and ratify 345, and to not consider any of the other 2,936 ballots (89% of the ballots) which they had placed on the “Consent” list, or more specifically the 2,576 votes (88% of the votes) preliminarily counted as VALID, which they never put separately on any list. (see also
Regardless of the reasoning used to evaluate any particular type of oddity, lapse or discrepancy—on the part of the delegates, proxies or volunteers—this approach guaranteed the outcome would not change. There was no need to apply any bias toward a candidate. And they knew very well that it didn’t matter WHOSE ballots they invalidated, or WHY, so long as they didn’t invalidate very many of them, so as to not change the final outcome (that is, who won the election).
So it was easy to give the impression of impartiality, an to seriously contemplate each of the various improprieties, all the while evincing a concern for “fairness” and compassion for the “intent” of the voter, so long as it resulted in a minimum of reversals. A mere preference toward validating as many ballots as possible would do the trick.
et 2,976 ballots (about 89% of the ballots) constituting (about 88% of the votes) were winnowed out as NOT worthy of further consideration, based on those BaRFs, which I feel I have adequately debunked. A 2.2% error rate is not acceptable when the original vote margin was about 2.1%. The winnowing process is flawed, therefore any conclusions based on that winnowing are unreliable. The dog-and-pony show on the 22nd involved 355 ballots chosen, by an unreliable method, for further review.
The CRC also makes illogical statements such as that finding one accusation of an altered ballot was actually an incorrect BaRF proves that no “tampering” occurred (of course “tampering” was never defined), rather than proving only that the BaRFs are fallible. Which is my point. Plus, the CRC conflate totals of “valid votes” with non-votes, and misrepresent the number of “Signature Issues” and of “Other Issues” that they “considered” (that is, they can’t even count how many names are on the lists they published).
And with amazing hubris, they also state unsubstantiated conjectures (more like apologetics) about their own fairness, as if they were “findings”. Case in point: the bizarre claim that the fact that they themselves found a nearly proportionate amount of Ellis ballots to question, as Ellis found Bauman ballots to question, proves the election process and the CRC process must have been unbiased. Given that they STARTED with 200+ complaints from Ellis, the CRC’s ability to find a couple hundred Ellis votes to challenge looks much more like an indication of bias than a vindication. And many of those were Ellis ballots being questioned for signature issues, and of course, the CRC are not signature experts, but it’s easy for anybody to question a signature.
July 10th, 2017
Compliance Review Commission
California Democratic Party
To Whom It May Concern:
I am writing to the Compliance Review Commission (“CRC”) as a dues paying member of the DSCC to the California Democratic Party with concerns I have about the CRC’s review process of the recent Chair Election.
As soon as the materials were made available to the public, I immediately began analyzing them to better identify issues to address on behalf of myself and other delegates to the California Democratic Party. Given that the CRC provided ballot metadata in the form of PDF scans of the Ballot Review Sheets, which I then had to transcribe manually into spreadsheets for this analysis, five days was insufficient to perform the analysis perfectly to my level of satisfaction. Furthermore, the short notice, lack of explicit directions on how to check ballots and the limited amount of time (especially during summer when people are vacationing) given to the delegates to review and properly respond to the CRC is insufficient; there should be at a minimum seven calendar days for everyone to have a fair opportunity to challenge any results of this review.
That said, with the time allotted, I was able to identify numerous inconsistencies in both the processes and the enforcement of the processes associated with credentialing and balloting that cause me great concern. Examples include inconsistencies around if staff provided credentials to people who had not paid or waived their dues or registration fees, inconsistencies in staff’s directions to delegates which led to some ballots being flagged for review while others were not, inconsistencies with which people were allowed to pick up credentials on behalf of others, and so on, as described in detail below.
In summary, there are at least two primary actions resulting from this analysis. First, a number of observable problematic patterns can easily be improved going forward. I plan to work with delegates to draft rules changes for checks and balances regarding credentialing so that the same problems arising from this election do not happen again.
Secondly, and most importantly for the matter of the Chair’s election, the CRC’s review demonstrates the limitations of the Commission’s undertaking: it comes woefully short of a thorough and transparent independent audit, which a significant number of delegates, including myself, want to see. The Commission should be commended for undertaking this review, which has been a huge endeavor. But there are significant problems arising from the Commission’s review: from the many variances in how the Ballot Review Forms were filled out, including differences in which sections were marked or unmarked, confusing notes and inconsistencies with the handling of ballots, and contrasting (even occasionally contradictory) handling of the same type of issues from ballot to ballot. Furthermore, there are major gaps in which information was reviewed and transcribed in the Ballot Review Forms, including a lack of verification of when payment or waiver of dues/registration were made, no information on whether County Central Committee lists were submitted in a timely fashion according to the bylaws of the respective counties, and whether delegate and proxy voter eligibility/registration have been verified in any consistent fashion.
Lastly, due to the composition of the Compliance Review Commission, I must also object to the lack of impartiality of the Commission members. Of the Commission’s six individual members, five voted for Eric Bauman for Chair. As participants in the election that they are evaluating, the Commission inherently increases the risk that their unconscious biases can affect their decisions and recommendations around the Chair’s race, including balloting and credentialing evaluations.
Given these findings, the Commission’s review is no substitute for an independent audit.
I submit to the Commission the following analysis of the CRC Ballot Review as official testimony. After identifying the following issues, unless the CRC is willing to address the issues outlined herein within its review, it is even more evident that the Party must perform an independent, forensic audit of the Chair Election results. On behalf of the delegates calling for an independent audit, we feel we have exhausted all remedies via the administrative due process made available to us by the CRC and have found it to be insufficient. The only available remedy that will bring the transparency and comfort to the aggrieved delegates in the party is to perform an independent forensic audit.
Katrina Bergstrom, Esq.
DSCC Delegate, AD43
Issues with “Valid” Ballots Review Sheets:
Record of Timely Payment Issues:
- 13 Review Sheets did not have “Record of Timely Payment/Waiver” Checked – were these not verified for payment or waiver?
Questionable Validity of Ballot:
- 1 Review Sheet did not have “valid/invalid ballot” at bottom of sheet checked
Questionable Validity of Proxies:
- 7 Review Sheets had “Certificate of Membership” to County Central Committees signed by Board members – There is no indication of timestamp of receipt of these and verification they were in compliance with county central committee bylaws.
- Were they attached after the convention?
- Were the County Bylaws followed in each instance?
- 1 Review Sheet is marked as “CRC Review” as proxy is not on county roster – also, there is no indication of who this person voted for (it’s not filled out) – this should not count if the county roster was not received before the Convention or whether they were in compliance with county central committee bylaws:
- 2 Review Sheets were marked that they did not have proxy forms in the proxy section at top – but are marked as “valid” at bottom:
- Whereas the following have been marked “CRC Review” due to missing Proxy Form
- It should not be permitted to accept copies of proxy forms after the fact, if they were not received at the credential table, then they missed the deadline and the vote should not count.
- Also, why were delegates given their credentials if they did not provide their proxy forms to the credential volunteers?
- 1 Review Sheet was marked as having “Duplicate of Proxy Form Pre-Approved Binder (Stamped)” but marked “valid” at bottom of sheet.
- Same issue was flagged for CRC review – why are these flagged and not the above?
- 1 Review Sheet was marked as “Valid” but the notes indicate that proxy is not on DSCC – why is this valid – Notes are very unclear.
- 1 Review Sheet indicated that on a valid ballot, although there was a proxy, that proxy didn’t vote, the original delegate voted. The proxy had been marked “invalid” due to not being on roster, and not being able to look up registration. How does the CRC verify the delegate was actually the one casting the ballot and not the invalid proxy?
- 1 Review sheet marked that the delegate who did not appear to attend the convention (no sign-ins), but is marked as voting (chair is blank), did not pay her registration but paid her dues, did the delegate vote for other positions? If so, why is she marked as valid – shouldn’t she be marked as “invalid” regardless of whether she voted for chair?
- 1 Review sheet is marked that the proxy signed in at registration, but did not sign for election or ballot, but ultimately voted. The chair is marked as “blank” – if she voted for other positions, why is she marked as “valid” shouldn’t she be marked as “invalid” regardless of whether she voted for chair?
- 1 Review sheet is marked as “No (No Blank)” with nothing else marked. However, an updated review sheet for this delegate indicates that he did in fact vote, but there are no notes to explain the sudden appearance of his ballot. Why is there no explanation?
- 3 Review Sheets that are marked as being “valid”, are missing information on whether they voted for chair or not. Two of these (Rodolfo and Tommy) are marked as having voted. Shouldn’t these be marked at “invalid”?
Issues with “CRC Review” Ballots Review Sheets:
- There are numerous Review Sheets which appear to be elected officials who have had their ballots picked up by staff. As I have been told (though it is not a written bylaw), this is a long-standing accommodation for elected officials and as long as the staffer leaves a card of the official, then it is permitted. HOWEVER, there appears to be a number of other delegates who are connected to the officials who have been flagged as having signature mis-matches. The courtesy extended to elected officials should not also be extended to appointees of the elected officials. These ballots should be found to be invalid. Below are some examples of this.
- In my review I found at least 19 PLEO appointed delegates who have been flagged for signature review by CRC. Due to the fact that there is a strong possibility these delegates had their credentials picked up by staffers for their PLEO appointer, these particular delegates’ signatures should have a higher scrutiny, and in fact they should be required to sign affidavits under penalty of perjury that they actually picked up their own credentials.
- For all of the elected official that are listed under CRC review, all of ballot review forms have “Signature Match” marked, but a handful of them have no signature for registration. There is no indication on the form of whether a staffer signed the name of the elected or themselves. This should be clear on all the forms.
- The following are marked “no signature” at registration:
- The following are marked as having signature at registration, there is no indication whether the staffer who picked up these ballots signed their own names or the names of the official:
- There are three ballot reviews for elected officials/appointees that are not marked under “record of timely payment/waiver” – There should already be record of these and the review sheet should be marked to indicate whether they have been received already. Considering Alex Padilla’s ballot has been marked invalid due to non-payment of dues & registration, these should be also.
- 1 Ballot Review Form indicates that the marking of the chair vote is unclear. As the instructions on the ballot clearly indicate that you must either make a “check mark” or an “x”, this ballot should be marked “invalid”:
- There are 18 Ballots marked as having no “registration signature” and no “signature match” – however there is only one (Patricia Acosta) that states that “volunteer failed to get signature” – why is that one review given such deference over the others? This shows an inconsistency in the review process.
- There is one ballot that indicates that although the signatures are substantially similar, the delegate submitted testimony that he signed for his wife’s ballot at the direction of the volunteer. There are two things that are brought to light here. 1. If the CRC is charged with reviewing signatures, and they were unable to find a difference in this one, then there could very likely be others that the CRC is unable to discern. 2. The volunteers at the Convention were under trained and were providing false direction to delegates – the process needs to be tightened up and all volunteers should have clear direction on the whole process and the importance of requiring signatures by the actual delegate themselves.
- There are 13 Ballot Review Forms listed as “CRC Review” who do not have anything checked under “Record of Timely Payment/Waiver” – although it doesn’t appear that they are being reviewed for that since they are all also marked “Signature Match” – Were these delegates checked for timely payment or waiver?
- If they had not paid, why did the credential volunteers give out their credentials? As I understand it, the credentials volunteers had a list of our names and whether we had paid our dues and registration. This seems like a commonsense rule that was not followed.
- Again, this shows a major inconsistency in the CRC review as far as checking all items on the ballot review form.
- There are 5 Ballot Review Forms that are marked “CRC Review” that either indicate that a ballot was not cast, or do not indicate anything (the boxes are unchecked) – there is no indication as to why the CRC are reviewing these ballots. If they did not cast a ballot, then they should be marked as “valid no vote”:
Issues with “Invalid” Ballots Review Sheets:
- There is 1 Ballot Review Form that is marked “invalid” but there is no clear reason why. The box for “Timely receipt of payment/waiver” is not marked, but there are no notes associated with it.
- There are 24 Ballot Review Forms that are marked “invalid” for having no record of timely payment or waiver. The last names of the delegates range widely so it does not appear to be an insulated incident of one volunteer who didn’t know. How did so many volunteers not know that they should not provide credentials if the delegate is marked as having not paid or waived? Not only did this cause many invalidated ballots, but the Party was damaged financially from not collecting registration to offset the cost of the convention.
- There are 10 Ballot Review Forms marked “Invalid” in which proxies were not on County Rosters. However, subsequent updates provided by the CRC indicate that 5 of these are now valid because “County Roster was received Electronically but not in folder”.
- Why were these immediately marked as “invalid” by the CRC if there was other ballots in a similar situation (see above) where they were also not listed on County Roster but were marked “CRC Review” instead of “Invalid”?
- There is no indication provided in these updates that tell us what the timestamp of the receipt of these updated rosters happened. If they were not received before the convention or if they were not in compliance with county central committee bylaws, then these should not be validated.
- There are 2 Ballot Review Forms that were marked “invalid” but do not have enough info filled out on them to determine who they voted for. These should be filled out completely.
Issues with “No Vote” or “No Chair Vote” Ballots Review Sheets:
- In my review, I found 331 Ballot Review Sheets that were not marked “invalid” or “CRC Review” and indicated no vote was cast.
- 310 are marked, “No (With Blank)”
- 21 are marked “No (No Blank)”
- There is no clear indication what the difference between these two ballot types are.
- There are varying levels of Ballot Review Form completeness in this category with many sections being left unchecked. For example:
- 237 have nothing checked under “Record of Timely Payment/Waiver”
- 75 are marked “no” under “Record of Timely Payment/Waiver”
- 19 are marked “yes “ under “Record of Timely Payment/Waiver”
- There is no indication as to why some were verified for payment/waivers and some were not, which shows inconsistencies with the CRC Review.
- There are 21 Ballot Review Forms where a ballot was cast, but no vote was made for Chair. 6 of these either were marked as having no signature match, or no record of timely payment/waiver. However, they are not marked invalid. Even if they didn’t vote for chair, these ballots should be marked invalid.