Wednesday, May 26, 2010

7 Ways Instant Runoff Voting Undermines North Carolina Verified Voting law

Instant runoff voting undermines key principles of North Carolina's nationally respected verified voting law, known as The Public Confidence in Election Act. This law was passed in August 2005 after hard work by activists and citizens from around the state. This law has done much to restore the integrity in our elections caused by faulty paperless voting machines and lack of standards for voting systems and vendors. IRV threatens key tenets of that law.

Why does IRV damage North Carolina Election Transparency Laws?

Procedures to implement instant runoff voting violate several parts of North Carolina election transparency laws, and key principles of election integrity. This is mostly due to the complexity of counting Instant Runoff Voting. IRV is complex to count because it is not additive. There is no certified software to tally IRV with, so uncertified "work-arounds" have been instituted. In touchscreen counties the NC SBoE says for purposes of IRV, the manual hand to eye count of the voter verified paper trail no longer rules, instead officials will substitute a single spreadsheet in for "manual counting". Because IRV is not additive, votes have to be centrally tallied, i.e hauled away from where cast to be counted at another location. IRV votes are not counted or reported on election night either. Worse, some IRV votes are never counted or reported.

Tallying procedures for instant runoff voting are error prone and complex. With optical scan - there are 2 ways to tally IRV on optical scan - 1) by hand, or 2) by running the ballots through the machines 4 times with machines reprogrammed before each run. Both are complex. But with touchscreens, the only certified option would be a manual tally of the paper trail. The State Board of Elections procedures are to use a spreadsheet and not manually tally the paper trail. So for IRV, the paper trail is not used for anything.

1. IRV undermines North Carolina standards for voting systems and vendors. There is no certified software to tally Instant runoff voting on North Carolina voting machines. Manual counting of IRV on touchscreen VVPATS would be difficult, so an illegal work around was devised. The NC State Board of Elections devised a complicated uncertified workaround to copy the vote data over to a spreadsheet. This error prone work around requires following 5 pages of single spaced instructions.

2. Audits and Recounts are undermined: The SBOE IRV touchscreen tallying procedure undermines the hard fought for 2005 legal requirement to use voter verified paper trails for manual audits and recounts.

The SBoE IRV procedures replace the Voter Verified Paper Trail with a "Spreadsheet":
In regular elections, when there is a recount or audit, volunteers examine the voter verified paper trail. But with IRV, in recounting or auditing in touchscreen jurisdictions, the State Board of Elections SUBSTITUTES an Excel spreadsheet prepared by officials - for the "voter verified paper trail".

3. IRV undermines the voter verified paper trail - for IRV in touchscreen counties, the voter verified paper trail will not be the count of record in the case of audits or recounts. A spreadsheet by an election official will usurp the paper trail.

Excerpt from SBE Guidelines for IRV:

Instant Runoff Voting:
Goals, Standards and Criteria for Implementation and Evaluation
Approved by the North Carolina State Board of Elections on January 15, 2009

"The 2006 Act specifies that the SBE may make modifications as it deems necessary .....
For jurisdictions located in counties using DRE machines, the first round of counting similarly can be conducted using the DRE’s regular automated counting but as a practical matter subsequent rounds, if necessary, cannot be counted directly by the DREs themselves or by the central tabulation software. For that reason the staff of the SBE have developed methods by which the vote information from the DRE is transferred to an Excel or other approved spreadsheet, validated for transfer accuracy, and then sorted and counted, either through the regular Excel functions or by hand-to-eye from the Excel spreadsheet. Any interested person may obtain a copy of the algorithm and spreadsheet.
Hand-to-eye counting is, by statute (G.S. 163-165.7), a certified voting system where the hand-to-eye count is from paper ballots marked by the voter, or marked by machine and verified by the voter prior to the ballots being cast. For that reason, the second round of voting where OS machines are used clearly involves a certified voting system. The second round of counting where DRE machines are used, however, involves a hand-to-eye count from an Excel spreadsheet or an automated count through regular Excel functions. In either case, the Excel software only serves the purpose of an electronic adding machine to total the votes.

4. NC law requires that votes be counted where cast. But with IRV, ballots are counted away from where they were cast. In Cary, NC - the 2nd and 3rd choice votes for the "instant runoff" were not counted on election night. Instead, they were carried away from where they were cast and then counted at a later date. Below is section of law that IRV pilots violate:

§ 163-182.2. Initial counting of official ballots . (a) The initial counting of official ballots shall be conducted according to the following principles:(1) Vote counting at the precinct shall occur immediately after the polls close and shall be continuous until completed.

Violating this law puts those choices at risk of tampering after being hauled away from the polling places and put into storage.

5. There is no Election night reporting for voters' second and third choices. There are no election night reports/poll tapes for IRV results because the machines cannot count the 2nd and 3rd choices, and officials do not themselves count the 2nd and 3rd choices on election night. North Carolina's voting machines are incapable of doing such reporting, The only way to account for the voters choices and secure them against fraud would be to count all second and third choices on election night at the polling places and create a manual report of that data. Election night reporting helps deter fraud from occurring while ballots are in transit.

6. Incomplete vote data. We don't get report of all votes cast in IRV elections.
In Cary's Oct 2007 IRV election, only partial data was reported for the District B contest where voters second and third choices were ultimately counted. None of the raw vote data was reported for the other IRV contests that had a winner in the first round. Without all raw vote data, i.e the tallies for all choices, whether they were needed to ascertain a winner - we cannot ascertain if this election was "non monotonic", i.e voters hurt their first choice by voting for them. Campaigns cannot see where their efforts succeeded or failed in campaigning and even cross endorsing as IRV advocates promote.

7. In Cary's 2007 IRV election, provisional ballots were not counted until after the 2nd and 3rd choices were counted, and supposedly "added" back in. Since IRV is not "additive", it is not clear how these votes could possibly be added back in without doing a complete recount. In SBoE new procedures, provisional ballots will only be counted "if needed".

Visit this link to sign up for email updates:

No comments: