Frequently Asked Questions
Referendum 71 Signature Check
Nick Handy
Director of Elections
Office of Secretary of State
August 13, 2009
Question: What is Referendum 71?
Response: Referendum 71 is a proposed ballot measure that seeks a statewide vote on Senate Bill 5688. SB 5688 is the “everything but marriage” law that expands rights for couples registered with the Secretary of State’s domestic partnership registry.
Engrossed Second Substitute Senate Bill 5688 was passed by the legislature in 2009.
The state referendum process, authorized in the state constitution, allows sponsors to submit a referendum petition that refers a bill passed by the legislature to a vote of the people.
In this case, if a requisite number of valid signatures are submitted, Referendum 71 will appear in the 2009 November general election for a vote of the people.
The law passed by the legislature is currently on hold pending a signature check by the Secretary of State’s Office and the outcome of the referendum.
Question: How many signatures are required to put a referendum on the ballot?
Response: Washington State law requires that 120,577 valid signatures be submitted to the Secretary of State by July 25, 2009 to get on the 2009 general election ballot in November 2009.
The Washington State Constitution established the required number of signatures for ballot measures. Referendum sponsors must submit a number of signatures at least equal to 4% of the votes cast for the office of Governor in the most recent gubernatorial election in Washington State.
Question: How many signatures did sponsors of Referendum 71 submit?
Response: Sponsors of Referendum 71 submitted approximately 137,689 signatures on July 25, 2009.
Question: The Office of Secretary of State recently changed the manner in which information is being provided to the public. Could you explain why these changes were made?
Response: Because of the high interest in this petition, and the obvious fact that the signature check would be very close, the Office of Secretary of State decided early that it wanted to make as much information available as soon as possible to the media and interested parties.
As a result of this decision, the office began posting on its website at the close of each business day the number of signatures checked and rejected for that day.
This information was posted mid-stream, before the master checkers had completed researching the registrations and reviewing rejected signatures.
As a result of this approach, numbers changed dramatically after the full research and review was complete, but those subsequent changes were not reflected in latter posts of information. The resulting data was confusing and created an impression of uncertainty about the process. This confusion and uncertainty was expressed in editorial articles and blog postings.
As a result, the Office of Secretary of State changed the method of reporting information to the public.
The most significant change is that information is not being reported mid-stream but is being reported after both checkers and master checkers have completed the research and signature check process on a bound volume of petition sheets.
There are two relatively small categories of signatures that will remain fluid, and could change after these daily releases:
- There are registrations in which the signature image on the state voter registration database may be of poor quality, or was not transmitted properly from the county voter registration system to the state voter registration system. In this very small number of situations, the state requests the county to provide a better image of the signature. The state will continue to receive those improved signature images through the verification process.
- There are registrations submitted recently that may not appear in the database being used by most checkers. When checkers and master checkers search for a signer’s name to ascertain whether the person is registered to vote, they actually search on a copy of the statewide voter registration database; they do not actually search on the statewide voter registration database itself. For security purposes, and to maintain the accuracy of the voter registration rolls, most checkers are not provided password access to the live statewide voter registration database. The copy that checkers are using is an extract that was pulled at the end of June in order to conduct the signature verification for Initiative 1033, which occurred at the beginning of July. Consequently, any registration applications that were submitted after this data extract will not be available to the checkers researching whether a signer is registered to vote. The state will use the live statewide voter registration database to confirm whether a signer registered more recently.
In the current form of reporting the progress of the verification, data will be reported by volumes completed, rather than by signatures initially checked in a day.
Signature petitions are randomly bound in volumes of 15 petition sheets per volume. For this signature check, 623 volumes have been created. The number of signatures in each volume varies because not all petition sheets have the maximum of 20 signatures per page.
Once a volume has been both checked and master checked, the Office of Secretary of State is releasing the signature results by volume. This reporting will occur once daily, typically in the afternoon, in spreadsheet format that will allow those closely observing the process to run their own calculations. This method of reporting provides more information because it reflects the final outcome, rather than a mid-process snapshot.
Question: Who verifies signatures on a referendum petition to ensure the proper number of valid signatures have been submitted?
Response: For over 50 years the Office of Secretary of State has been charged by statute with verifying signatures on initiative and referendum petitions.
Question: Is the Secretary of State authorized to perform random checks of signatures or must every signature be checked?
Response: The legislature has authorized the Secretary of State to establish, by rule, a random sample method to perform signature checks of initiatives and referenda.
The Secretary has published rules establishing criteria for conducting a random sample signature check. The standards used by the Secretary are based upon a mathematical algorithm devised by a mathematics professor at the University of Washington.
Generally, the random sample method can be used when a ballot measure sponsor submits a total number of signatures that far exceeds the required minimum – 25% or more of the minimum required.
For example, sponsors of Initiative No. 1033 in 2009 submitted 315,000 signatures. For an initiative petition, 241,153 signatures are required. Since the number of signatures submitted far exceeded the number of signatures required, the algorithm authorized a random check of 3% of the signatures submitted.
Question: Is Referendum 71 a random check or a 100% check, and how often is a 100% check conducted?
Response: Referendum 71 is being conducted as a 100% check. This is because the sponsors only submitted 137,689 signatures where 120,577 signatures are required by law. The margin of excess signatures is so low, about 17,000 signatures, that the Office of the Secretary is not confident the random sample procedure will accurately conclude that there are sufficient valid signatures to certify the measure for the General Election ballot. Therefore, the office has determined that a 100% check is appropriate.
Since 1990, the Secretary of State has conducted 57 signature checks on initiatives and referenda petitions. Only four of these, including Referendum 71, have been 100% checks. The remaining 53 checks were completed using the random sample method.
Question: Can an initiative or referendum petition be rejected on a random check?
Response: No. An initiative or referendum petition may be approved on a random check but may not be rejected on a random check. Thus, if a petition fails the random check, a 100% check must be conducted before the petition may be rejected.
Question: Please describe the process that is used to verify signatures on a referendum petition?
Response: The Elections Division of the Secretary of State’s Office hires temporary workers to perform signature checking on referendum petitions, though many of these temporary workers have experience checking petitions from previous summers.
For Referendum 71, the Elections Division has hired approximately 30 signature checkers and 8 master checkers. The Elections Division is running two work shifts per day, from 7:30 am to 3:30 pm, and from 3:30 p.m. to 10 p.m.
Referendum petitions are bundled into volumes containing 15 petition sheets. Because the number of signatures on each petition sheet varies, the number of total signatures in each packet varies accordingly. Each volume has an assigned number.
A signature checker “checks-out” a volume and proceeds to review all the signatures in that volume. The signature checker works at a computer terminal. The checker enters identifying information for each signer to bring up that signer’s voter registration record. Once the voter registration record is identified, the checker compares the signature on the petition with the signature on the voter registration record.
Question: What are the reasons for rejecting a signature?
Response: Signatures are rejected based on one of four reasons:
- The signer does not appear to be registered to vote in Washington. A registration for the person was not found. These signatures will be rejected because a signer must be registered to vote in Washington for the signature to be valid. An initial checker may be unable to find the signer’s registration among the over 3.6 million people registered to vote. No date of birth is provided on the petition, some people do not print legibly, and some people provide a different name than their registration name. During a subsequent search, conducted by a more experienced master checker, a signer’s registration file may be found. In addition, a subsequent search on the live statewide voter registration database may reveal a registration recently submitted after the start of the I-1033 signature verification process.
- The signature on the petition cannot be compared to the signature on the state list of registered voters. The digital image of a signature in the statewide voter registration database may be of poor quality, or was not transmitted properly from the county voter registration system to the state voter registration system. In these cases, the Secretary of State works with the county where the signer is registered to vote to obtain a better image of the voter registration signature. Once a better signature image is obtained, it will be reviewed. Experience indicates that many of the signatures in this category are accepted. This category of signatures will remain fluid until the end of the verification process because the Secretary of State’s Office will continue to receive the images from county election offices until that time.
- The signature on the petition sheet does not match the signature in the person’s voter registration file. A signature that does not appear to match the signature on file must be reviewed by at least two checkers before it is rejected.
- The same signature appears more than once on the petition sheets. Only one signature from each registered voter may be accepted. Once a voter’s signature is accepted, any duplicate signatures on Referendum 71 petition sheets found after that are rejected. Experience indicates that this number will slowly increase as a full verification progresses and more signatures have been accepted.
Question: What is the role of master checkers?
Under policies established by the Office of Secretary of State, a signature may not be rejected by just one checker.
Any signature that has been rejected by a signature checker is also referred to a master checker. To be finally rejected, a signature must have been rejected by a signature checker and the rejection must be confirmed by the master checker.
Master checkers are experienced checkers who have worked in the initiative and referendum checking process in previous years. They are more experienced at searching the voter registration rolls to find registered voters.
As a result of this experience, master checkers are more effective at finding voter registration records.
The vast majority of situations in which a master checker would change the decision of a signature checker would be where the master checker is able to locate a signer’s voter registration record, after the signature checker could not find the record.
Difficult to find registrations might include, for example, a common name like John Smith, with a difficult to read address. Another may be a situation where a voter has moved and did not update his voter registration information with the County Auditor. A third situation might be where a voter has used a different first name than what appears on the voter registration record such as Meg instead of Margaret, Liz instead of Elizabeth, or Jack instead of John. Other situations may involve voters who have changed their name as a result of marriage or divorce. The petition requests the signer’s name, signature, and address; it does not request a date of birth or other identifying information that would assist in locating the signer’s registration. For example, there are over 32,000 people registered with the last name of Smith, and over 32,000 people registered with the last name of Johnson.
Master checkers are more experienced and skilled at searching the database for these difficult to find voters.
The more experienced checkers are also used to search the live database for recent registrations that were entered after June 19, after the data extract occurred for the summer initiative and referendum signature checks.
Question: Do master checkers double check approved signatures?
Response: Master checkers only review rejected signatures. Master checkers do not review approved signatures.
In the case of approved signatures, the signature checker has located the voter registration record and has compared the signature on the petition sheet with the signature on the voter registration record. These signatures either match or do not match. In close cases, the checker may request a supervisor or master checker to help with a close call. To recheck all approved signatures would nullify the vast majority of verification work already completed.
Question: What percentage of signatures submitted by Referendum 71 sponsors must be approved to meet the requirement of 120,577 valid signatures?
Response: Given that Referendum 71 sponsors submitted approximately 137,689 signatures, approximately 87.58% of these signatures must be found to be valid in order to confirm 120,577 valid signatures.
Stated in the converse, the Referendum 71 signature rejection rate must be 12.42% or lower.
Question: What are the typical approval or rejection rates that have been achieved by other initiative and referendum petitions
Response: The Office of Secretary of State has posted here the results of all initiative and referendum petitions processed in the last 20 years. The rejection rate on these petitions ranges from a low of 10% to a high of 25%. The average rejection rate is typically 15 to 18%. On Initiative No. 1033 approved just last month, the rejection rate was 12.14%.
Question: Where is the current signature verification information for Referendum 71 posted?
Response: The most current information is posted on the Elections Division website here. The rejection rate is expected to increase as more duplicates are found.
Question: How do duplicates factor into the process?
Response: Duplicates are probably the single most significant factor in a close signature check such as Referendum 71.
Once a representative sample of signatures has been checked, trends become established for the number of signers who are not registered or the number of signers whose signatures do not match. Once established, it would be unusual for these trends to change.
That is not the case with duplicates. When only 1,000 signatures have been checked, the next signature checked is compared against that pool of 1,000 signatures already checked. The likelihood of finding a duplicate in a pool of 1,000 signatures is relative low.
But, when 130,000 signatures have been checked, the likelihood of finding a duplicate signature in the much larger pool logarithmically increases.
Thus, as the pool of signatures checked increases, the likelihood of finding duplicates increases.
The result of this dynamic with duplicates is that the rejection rate that appears established early in the check is always expected to increase as the signature check progresses.
Question: Given the current rejection rate and the number of duplicates being found, is Referendum 71 expected to pass or fail?
Response: The policy of the Office of Secretary of State is to make all data available as promptly as possible, but not to project the ultimate outcome of a petition signature check.
The Office of Secretary of State can only say that the numbers are extremely close on Referendum 71.
Question: What role do observers play in the process?
Response: State law allows for interested parties to have observers at all stages of the signature check process.
Normal state standards would allow two observers per side. In this case, based on a request by one side and the unusually close nature of this petition, the Office of Secretary of State has authorized three observers per side.
On the Referendum 71 signature check, six observers, three from each side, are actively monitoring the process.
Observers may stand next to a signature checker or a master checker and review the signature checking process. Observers are not allowed to talk with signature checkers or master checkers, and may not write down or otherwise record the names of signers.
Observers may convey to the Office of Secretary of State supervisors any concerns and may also record the volume, page and line number of a signature decision for which an observer may not agree.
In the Referendum 71 situation, both sides have requested the Office of Secretary of State to review certain decisions and these requests are being accommodated.
Question: Is the Referendum 71 process in litigation?
Response: The Office of Secretary of State has received five public records requests for copies of the Referendum 71 petition sheets.
As a matter of state law and past practice, the Office of Secretary of State has always made these petitions sheets available to any person making a public records request. Because a digital image was taken of every petition sheet submitted, this information is available digitally on a CD disk and in other formats.
The copies of the petition sheets were made before the verification process began, and therefore do not include any notations on which signatures were later accepted or rejected.
A lawsuit has been filed in federal court to prevent the release of this information and a federal district court judge has entered a temporary restraining order preventing the Office of Secretary of State from releasing this information. A court date has been set in September for deciding these issues.
The five public records requests, and documents filed in the litigation, are posted on the Secretary of State website here.
Question: When will the Referendum 71 check be completed?
Response: The Office of Secretary of State projects that the signature check process will be completed by the end of August, 2009.