Summary

County Poll Workers: The Office of the Secretary of State Has Developed Statewide Guidelines, but County Training Programs Need Some Improvement September 2008 Report 2008‑106 C A L I F O R N I A S T A T E A U D I T O R The first five copies of each California State Auditor report are free. Additional copies are $3 each, payable by check or money order. You can obtain reports by contacting the Bureau of State Audits at the following address: California State Auditor Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, California 95814 916.445.0255 or TTY 916.445.0033 OR This report is also available on the World Wide Web http://www.bsa.ca.gov The California State Auditor is pleased to announce the availability of an on‑line subscription service. For information on how to subscribe, please contact the Information Technology Unit at 916.445.0255, ext. 456, or visit our Web site at www.bsa.ca.gov. Alternate format reports available upon request. Permission is granted to reproduce reports. For questions regarding the contents of this report, please contact Margarita Fernández, Chief of Public Affairs, at 916.445.0255. CALIFORNIA STATE AUDITOR Elaine M. Howle State Auditor Doug Cordiner B u r e a u o f S t a t e A u d i t s Chief Deputy 555 Capitol Mall, Suite 300 Sacramento, CA 95814 916.445.0255 916.327.0019 fax www.bsa.ca.gov September 18, 2008 2008-106 The Governor of California President pro Tempore of the Senate Speaker of the Assembly State Capitol Sacramento, California 95814 Dear Governor and Legislative Leaders: As requested by the Joint Legislative Audit Committee, the Bureau of State Audits presents its audit report concerning the county elections officials’ training of poll workers. Specifically, the report addresses the role of the Office of the Secretary of State (office) in providing guidelines to county elections officials for training poll workers and certain aspects of county poll worker training programs. Based on recommendations from a task force established by state law, in 2006 the office adopted poll worker training guidelines (training guidelines) that establish a minimum set of requirements that counties must meet. However, the guidelines do not cover the rights of voters who register to vote without declaring a political party affiliation (decline-to-state voters). The law does not require the office to update the training guidelines and it has not done so since issuing them in 2006. Nevertheless, senior management stated that the office plans to update the training guidelines but would need to convene a task force similar to the one used to develop the original guidelines. Though not required, the office has performed some monitoring of counties’ administration of elections through its Election Day Observation Program. During the February 2008 primary election the office visited 31 counties and afterwards shared its observations with each county to help identify how to strengthen their respective training programs. According to the deputy director of operations, future election observations will depend upon available resources and whether changes in the law require changes in polling place operations. The eight counties we reviewed substantially complied with the content of the office’s training guidelines when training their inspectors and other poll workers. However, some counties appeared to only partially train poll workers in certain areas, such as voters’ rights and the operation of voting machines. Some of the eight counties we reviewed had embraced noteworthy training practices such as optional workshops with more opportunities for hands-on training and online training, and some varied the content of their training to match the experience level of attendees to promote greater training attendance. However, not all counties required inspectors to attend training prior to working elections and many had difficulty providing complete and accurate lists of inspectors that received training prior to the February 2008 election. The eight counties we visited told us they use a variety of sources for collecting information to identify needed improvements in their poll worker training programs. Although all eight counties asserted that they used some variation of feedback to evaluate their training programs, none of the counties could clearly demonstrate how the information collected from the February 2008 election was summarized and used to make changes in their training programs for the June 2008 election. Most of the counties we reviewed discussed procedures for handling voter complaints in their poll worker training; however, the emphasis the counties placed on handling complaints varied, and many of the counties were unable to provide reliable data that described how they resolve complaints. Finally, all eight counties said their poll worker recruitment goals were based on projected voter turnout, projected absenteeism, and the need for multilingual workers, and many of the counties reported challenges in recruiting an adequate number of poll workers. Respectfully submitted, ELAINE M. HOWLE, CPA State Auditor County Poll Workers: The Office of the Secretary of State Has Developed Statewide Guidelines, but County Training Programs Need Some Improvement September 2008 Report 2008‑106 California State Auditor Report 2008-106 vii September 2008 Contents Summary 1 Introduction 7 Chapter 1 The Office of the Secretary of State and Other Sources Assist the Counties’ Elections Officials in Developing Training for Poll Workers 15 Recommendation 22 Chapter 2 Although Generally Consistent With Guidelines for Training Poll Workers, Most Counties’ Training Programs Need Improvement 25 Recommendations 47 Appendix Counties Generally Comply With Training Guidelines From the Office of the Secretary of State Relating to the Subjects Covered in the Inspector Training Class 51 Responses to the Audit California Secretary of State 53 Alameda County Registrar of Voters 55 Fresno County Registrar of Voters 59 California State Auditor’s Comments on the Response From Fresno County 61 Kings County Registrar of Voters 63 Los Angeles County Registrar of Voters 65 California State Auditor’s Comments on the Response From Los Angeles County 71 Orange County Registrar of Voters 73 California State Auditor’s Comments on the Response From Orange County 77 San Diego County Registrar of Voters 79 California State Auditor’s Comments on the Response From San Diego County 83 viii California State Auditor Report 2008-106 September 2008 Santa Clara County Registrar of Voters 85 California State Auditor’s Comments on the Response From Santa Clara County 91 Solano County Registrar of Voters 95 California State Auditor’s Comments on the Response From Solano County 103 California State Auditor Report 2008-106 1 September 2008 Summary Results in Brief Audit Highlights . . . To be able to cast a ballot in an election in the State of California Our review of county elections officials’ (State), eligible individuals must register to vote. In doing so, they training of poll workers revealed can declare an affiliation with one of six political parties certified by the following: the Office of the Secretary of State (office), thus becoming able to participate in that party’s nomination process, or they can register » In 2006 the Office of the Secretary of without a political party affiliation. State (office) adopted poll worker training guidelines (training guidelines), as The California Elections Code (Elections Code) makes county required by law. elections officials responsible for training poll workers. State law also requires the office to administer the provisions of the Elections » The law does not require the training Code and to ensure that elections are conducted efficiently and guidelines to be updated and the office that state election laws are followed. In 2003, the State Legislature has not done so since issuing the training (Legislature) enacted a law that required the office to establish a task guidelines in 2006. force to recommend uniform guidelines for training poll workers. The guidelines were to include certain topics, such as voters’ rights » The office’s senior management asserts and polling place operations. In 2006, as required by state law, the that although the law does not direct the office published the Poll Worker Training Guidelines 2006 (training office to monitor counties’ compliance guidelines), which reflects the work of the task force. The document with the training guidelines, the office was not intended to take the place of training materials or resources does conduct some observations of for poll workers; rather, it was to establish a minimum set of counties’ elections and shares the requirements that training sessions and materials developed by results of its findings with the counties the counties must meet and to set a standard against which county it observes. programs for poll workers should be measured. » The eight counties we reviewed The law does not require the training guidelines to be updated, and substantially complied with the content the office has not done so since issuing them in 2006. Nevertheless, of the training guidelines when training senior management at the office have expressed a desire to update the their inspectors, but some counties training guidelines and have acknowledged that to do so, the office appeared to only partially train poll would need to convene a task force similar to the one used to develop workers in certain areas. the original training guidelines. » Some counties employed noteworthy One subject not covered in the training guidelines is the rights of practices targeted toward providing poll voters who registered to vote without declaring a political party workers with added opportunities to affiliation (decline‑to‑state voters). The office’s senior management practice what they have learned. stated that in the February 2008 presidential primary election, many decline‑to‑state voters were confused about which political » Not all counties required inspectors parties’ candidates they could cast ballots for because only two of to attend training or were able to California’s six qualified political parties had authorized this type demonstrate they trained all inspectors of voter to cast ballots in their primaries. In addition, some news prior to the February 2008 election. agencies reported that poll workers gave unclear instructions to decline‑to‑state voters and that poll workers were unsure as to how continued on next page . . . much information they could volunteer to these voters. The office has taken steps to eliminate voter and poll worker confusion, such as emphasizing the rights of decline‑to‑state voters in its June 2008 Voter Information Guide. Currently, there are two bills before the 2 California State Auditor Report 2008-106 September 2008 Methods for Increasing Management » None of the counties could clearly Legislature that would require poll workers to inform decline‑to‑state Personnel Salaries demonstrate how the information voters that they may request a ballot for a political party that has collected from the February 2008 election adopted a party rule allowing these voters to vote that party’s ballot. • Merit salary increase program: Performance‑based was summarized and used to update their salary increases funded from a merit compensation pool training for the June 2008 election. In addition to its guidelines, the office has communicated training established annually by the chancellor’s office. information through periodic memorandums (memos) to county • Equity (market) increase program: Adjustments » Many of the counties were not elections officials, as well as through trainings and informational designed to address discrepancies in pay, both within able to provide reliable data that seminars conducted by the California Association of Clerks and and outside the university system, for comparable jobs. described how they resolved voter Election Officials (CACEO), an association of county elections • Reclassification: Salary increases resulting and poll worker complaints. officials. The office uses the memos as a means of communicating from changes in administrative classification that with county elections officials about election‑related topics. Of the reflect changed assignments. more than 650 memos the office issued between April 2006 and April 2008, we found that 11 seemed to have implications for poll worker training. Although not required to do so, the office performs limited monitoring of the poll worker training conducted by counties. The office’s senior management noted that although the law establishes the secretary of state as the chief elections officer it does not direct the office to track whether counties conform to the office’s guidelines when training poll workers or to develop regulations or policies surrounding poll worker training. However, the office does perform some monitoring of counties’ administration of elections through its Election Day Observation Program (observation program). Created in 2003, the observation program began as a poll‑monitoring program that focused on preventing issues such as long lines at polling places and the intimidation of voters. Subsequent election reviews have focused on how well counties were complying with federal election requirements. During the February 2008 primary election, the office staff visited 31 counties and afterward shared their observations with each county to help them identify ways to strengthen their respective poll worker training. The office performed a similar review in June 2008, and the office’s senior management stated that they plan to perform a review in November 2008 but are uncertain about the 2010 election cycle. According to the deputy director of operations, whether the observation program will continue in 2010 is dependent upon available resources and whether changes in the law require changes in polling place operations that dictate a need to observe how the counties are implementing those changes. Many of the eight counties we reviewed look to other sources of information, rather than the office when updating their training programs. Three of the eight counties we visited told us they do not believe they are required to follow the training guidelines. One county told us that it seldom reviews the training guidelines for current elections because the guidelines have not been updated. Seven of the eight mentioned using the CACEO or the California State Auditor Report 2008-106 3 September 2008 United States Election Assistance Commission (commission) for information to update their poll worker training programs. The Election Administration Research Center (center) at the University of California, Berkeley, is another organization that provides tools to counties for improving their training programs. The center released two reports summarizing its findings from surveys of poll workers that the center administered during the 2006 election cycle. The eight counties we reviewed substantially complied with the content of the office’s training guidelines when training poll workers, which consist of the inspectors who supervise polling places and the clerks who staff them. However, some counties appeared to only partially train poll workers in certain areas, such as voters’ rights to report illegal or fraudulent activity, sensitivity to multiple cultures, and the operation of voting machines. Additionally, some counties did not use suggested training methods, such as role‑playing for processing voters’ ballots and hands‑on training for teaching workers to operate voting machines. However, after encountering problems in the February 2008 primary election with ensuring the rights of decline‑to‑state voters, the eight counties whose training we observed all discussed the voting options available to these voters prior to the June 2008 election. In our review of eight counties, we observed some noteworthy training practices. Most of these practices seemed targeted toward providing poll workers with additional opportunities to practice what they have learned while also being sensitive to their time commitments. For example, we found that some counties offered training at various times and locations and tailored the content to the experience level of the attendees to promote greater training attendance. Others offered online training or optional workshops with opportunities for more hands‑on training just prior to the election. Not all of the counties we reviewed required inspectors to attend training prior to working elections or were able to provide data demonstrating that they trained all inspectors prior to the February 2008 election. Specifically, many counties had difficulty providing us complete and accurate lists of inspectors that received training. As a result, many counties in our sample cannot be certain that all these workers have the knowledge to efficiently administer elections. The elections officials from the eight counties we visited told us they use a variety of sources for collecting information to identify needed improvements in their poll worker training programs. These sources included post‑training feedback from poll workers, comments from instructors, postelection debriefing reports, analyses of voter complaints, and reviews of questions 4 California State Auditor Report 2008-106 September 2008 from poll workers on election day. Seven of the counties were able to provide at least some documentation of the information they collected. However, none could clearly demonstrate how the information collected from the February 2008 election was summarized and used to make changes in their training programs for the June 2008 election. At most, counties were able to provide postelection evaluation reports that described what needed to be changed in their training programs for poll workers; however, these reports did not link their conclusions from the data collected to the proposed changes to be made. As a result, we could not determine whether the counties in our sample effectively used the information they collected to improve their poll worker training. Under state law, voters have the right to ask poll workers and elections officials questions and register complaints about election procedures and to receive an answer or be directed to an appropriate elections official for an answer. Although most of the counties we reviewed discussed procedures for handling voter complaints in their poll worker training, the emphasis the counties placed on handling complaints varied. For example, Alameda County did not specifically discuss handling complaints; instead it used its training guide to inform poll workers that voters can call the office or county elections officials with complaints. In contrast, Orange County dedicated training time to discussing voters’ rights to complain about election issues and developed referral cards with the county’s voter hotline telephone number for poll workers to hand out to voters who complained at the polls. Los Angeles, San Diego, and Santa Clara counties developed similar referral cards. Although all eight counties told us they receive complaint calls from voters or poll workers on election day, most counties we visited were unable to provide information on how they resolved voter or poll worker complaints. Finally, to determine the number of poll workers to assign to each polling place, most of the county elections officials we reviewed used general poll worker recruitment goals, such as ensuring that each polling place has at least one inspector and three to four clerks. The counties we visited cited various factors when considering how many poll workers to recruit for election day, such as projected voter turnout, expected poll worker absenteeism, past experience with elections, and the need for multilingual poll workers. Many of the counties reported challenges in recruiting an adequate number of poll workers, and they stated that they relied on practices such as expanding recruiting at schools and hiring a reserve of workers to make up for absenteeism among polling place workers on election day. California State Auditor Report 2008-106 5 September 2008 Recommendations The Legislature should consider amending the Elections Code to explicitly direct the office to periodically update its poll worker training guidelines and to monitor county adherence to these standards. In the interim, the office should continue with its plans to update its training guidelines and incorporate new guidance on the proper handling of decline‑to‑state voters. Finally, to the extent feasible, the office should continue its efforts to monitor county adherence to its guidelines through its observation program. To ensure that poll worker training programs conform with the office’s guidelines, county elections officials should review the content of their programs, ensuring their training fully covers topics such as voter complaint procedures, preventing voter intimidation, and issues pertaining to a culturally diverse electorate. To improve poll workers’ willingness to attend training and their ability to retain the lessons learned, county elections officials should consider implementing the following practices: • Maximize the number of training sessions scheduled for poll workers while also offering the training at multiple locations with different start times to better accommodate poll workers’ other time commitments. Also, providing condensed training tailored to experienced poll workers may entice greater attendance, while more extensive training can be reserved for new poll workers. • Offer poll workers an opportunity to reinforce what they learned in class through the use of online supplemental training material. Such an online program might include practice quizzes on election‑day procedures, examples of the election materials to be used, and reference materials provided at training. County elections officials might also consider providing podcasts that emphasize critical aspects of poll worker training. • Provide optional workshops giving poll workers additional opportunities to practice what they learned and to get hands‑on experience in the use of election‑day supplies and voting equipment. County elections officials might consider providing these workshops 0n the days immediately before an election to maximize poll worker confidence and retention of information. 6 California State Auditor Report 2008-106 September 2008 To better ensure that county elections officials provide knowledgeable inspectors to serve voters, counties should take steps to ensure that all inspectors receive training. Steps that counties might take to achieve this goal include: • Compiling accurate lists of inspectors who have attended training while informing inspectors who did not go through training that they cannot serve as inspectors. • Recruiting reserve poll workers who have gone through inspector training to be deployed, as necessary, to polling places where the assigned inspectors did not receive the required training. To better ensure that training programs for poll workers are effectively evaluated and needed improvements identified, county elections officials should consider taking steps to track voter complaints and poll worker questions that are received during an election, evaluate whether such comments suggest ways to improve their training programs, and implement those improvements. Agency and County Comments Overall, the office concurs with our recommendations as they relate to its operations. However, the eight counties we reviewed for the audit vary in their agreement with our findings and recommendations. The office’s and counties’ responses begin on page 53. California State Auditor Report 2008-106 7 September 2008 Introduction Background Eligible voters in the State of California (State) must register to vote in elections, and they can do so with their county elections officials or through certain designated public agencies, including the California Department of Motor Vehicles. Voters wishing to participate in a political party’s nominating process can register with one of the six qualified political parties certified by the Office of the Secretary of State (office). Doing so enables them to vote in that party’s primary election, in which candidates are chosen to run for office in the ensuing general election. Alternatively, voters may decide not to affiliate with a political party. During the February 2008 presidential primary election, about 20 percent of registered voters declined to state an affiliation with a political party (decline‑to‑state voters). Figure 1 lists the party affiliations of Californians registered for the February 2008 election. Figure 1 Number of California Voters Registered in Each Party for the February 2008 Presidential Primary Election Other—690,758 (4.40%)* Decline to state— 3,043,164 (19.37%) Democratic— 6,749,406 (42.95%) Republican— 5,229,425 (33.28%) Source: The Office of the Secretary of State’s report titled Historical Voter Registration and Voter Participation in Statewide Primary Elections—1910-2008. * American Independent Party, Peace and Freedom Party, Green Party, and Libertarian Party. In a primary election, a voter’s party affiliation determines the slate of candidates that appears on the ballot he or she receives. The State currently has a modified closed primary system, the result of legislation passed in 2001 that permits decline‑to‑state voters 8 California State Auditor Report 2008-106 September 2008 to participate in a primary election if an individual party’s rules authorize them to do so. Political parties have until 135 days before a primary election to provide written notice to the office indicating the adoption of a rule allowing decline‑to‑state voters to vote in that party’s primary. If a decline‑to‑state voter does not request a ballot for one of these parties, the voter will receive a nonpartisan ballot listing only ballot measures and the names of candidates for nonpartisan offices. In the February 2008 primary election, the Democratic and American Independent parties allowed decline‑to‑state voters to participate in their primaries. After the primary election, the winning candidate from each party moves to the general election, in which all voters, regardless of their political party registration, are allowed to vote for any candidate on the ballot. The Role of the Office of the Secretary of State in the Election Process California law makes the office responsible for administering the provisions of the California Elections Code (Elections Code) and requires the office to see that elections are conducted efficiently and the State’s election laws are enforced. If the office concludes that the appropriate officials are not enforcing state election laws, the office must call the violation to the attention of the district attorney of the county or to the State’s attorney general. The office is responsible for various activities related to the administration of elections in California, including overseeing all federal and state elections in the State. In every statewide election, the office prepares voter information pamphlets in seven languages. As the chief elections officer, the secretary of state directs her staff to test and certify all voting equipment for security, accuracy, reliability, and accessibility to ensure that polling officials count every vote exactly as it was cast. The office also is responsible for enforcing election laws and campaign disclosure requirements, maintaining a statewide database of all registered voters, certifying the official lists of candidates for elections, tracking and certifying ballot initiatives, compiling election returns and certifying election results, educating California citizens about their voting rights, and promoting voter registration and participation. In 2003, the Legislature passed a bill that required the office to establish a task force to make recommendations for uniform guidelines for training poll workers. The enacted law also requires that the guidelines, at a minimum, include the topics listed in the California State Auditor Report 2008-106 9 September 2008 text box. In 2006, the office issued Poll Worker Training Guidelines 2006 (training guidelines), a Topics that the guidelines for training poll workers must address, per state law: publication that reflects the task force’s work. The office intended the training guidelines to provide a • The rights of voters—including rights pertaining to starting point for counties’ training programs for linguistic minorities, the disabled, and protected classes poll workers. It did not intend for the document to under the federal Voting Rights Act. take the place of training materials or resources for • Election challenge procedures—including challenging poll workers; rather, the guidelines were meant to precinct administrator misconduct. establish a minimum set of requirements that training sessions and materials for poll workers • Operation of a jurisdiction’s voting system—including, must meet, and to set standards by which counties but not limited to, information on modernized voting can measure their programs. systems and proper tabulation procedures. • Poll hours and procedures for the opening and closing of polling locations on election day—including County Elections Officials’ Responsibilities for information on how to ensure that voters who arrive Conducting Elections and for Hiring and Training at a polling location before it closes are allowed to cast Poll Workers a ballot. • Relevant election laws—including any other The Elections Code makes county elections officials information that will help the inspector carry out his or responsible for training poll workers who serve her duties. as inspectors in polling places. County elections • Cultural competency—including knowledge of officials can be county clerks, city clerks, the diverse cultures and languages that poll workers may registrar of voters, or election supervisors who have encounter during the course of an election and the jurisdiction over elections within any county, city, or appropriate skills for working with the electorate. district within the State. County elections officials are • Knowledge of issues confronting voters who have responsible for creating election precincts, or distinct disabilities—including information regarding access geographical territories, within their jurisdictions barriers and the need for reasonable accommodations. and for designating a polling place for each precinct, which may have no more than 1,000 registered • Knowledge of procedures regarding provisional voters. County elections officials also appoint voting—including information about when poll workers can require voters to cast provisional ballots. precinct boards, or groups of poll workers, for each precinct, and the officials determine the Source: California Elections Code, Section 12309.5. composition of those boards. In general, a precinct board consisting of at least one inspector and two clerks must staff each precinct on election day. After creating the precincts, county elections officials must then train the precinct inspectors. State law requires only inspectors to receive training. In addition, the Elections Code allows the counties to hire students who meet certain requirements to work at polling places to increase the work force available to the counties and provide students with a greater awareness of the election process and the rights and responsibilities of voters. According to the training guidelines, county elections officials should train poll workers so that they understand that their role is to assist every qualified voter to cast a ballot and to make certain that each ballot is safely secured until it can be counted. Poll workers should also ensure that the rights of everyone seeking to vote are protected. In addition, poll workers should be familiar 10 California State Auditor Report 2008-106 September 2008 with the rights of voters, be trained in cultural sensitivity, know how and when to assist voters with disabilities or other specific needs, and know their responsibilities and the limits relating to them. The California Association of Clerks and Election Officials Assists With Poll Worker Training The California Association of Clerks and Election Officials (CACEO) is an organization whose purpose is to promote a high standard of administrative ability and efficiency in the county personnel who conduct the administrative affairs for the departments of county clerk, clerk of the board of supervisors, and registrar of voters. Another goal of the CACEO is to promote uniform practices and procedures in these departments’ application of their responsibilities. Principal membership in the CACEO is reserved for the county clerk, the registrar of voters, and the clerk of the board of supervisors in those counties having one or more county officers holding these titles. According to the CACEO’s president, all counties currently participate in the CACEO. Seven of the eight counties in our sample mentioned using the CACEO as a resource for updating their training programs for poll workers. The CACEO administers the California Professional Election Administration Credential Program (credential program), which certifies county elections officials. The credential program consists of 10 courses, and those who complete the entire curriculum earn a credential. CACEO leadership stated that the organization recently paired the credential program with the CACEO’s New Law Conference to help train more county elections officials. According to the CACEO leadership, the credential program has trained 170 CACEO members. According to its president, the CACEO also holds two conferences annually and uses all‑county conference calls to inform county elections officials of major issues and best practices. The credential program and the CACEO’s Summer Institute training program (summer institute) instruct elections officials on election‑related issues. According to CACEO leadership, each session of the summer institute for CACEO members covers a single topic in depth, such as candidate services, poll worker training, and absentee ballots. California State Auditor Report 2008-106 11 September 2008 Funding From the Help America Vote Act of 2002 Provides Minimal Training Assistance for the State’s Poll Workers According to the United States Election Assistance Commission (commission), a body created by the federal Help America Vote Act (HAVA) of 2002, the legislation made various reforms to the nation’s voting process. HAVA established the commission to help states comply with HAVA by serving as a national clearinghouse and resource for information about administering federal elections. HAVA addresses improvements to voting systems and voter access issues that were identified following the 2000 election. It also creates new mandatory minimum standards for states to follow in several key areas of election administration. The law provides funding to help states comply with HAVA requirements. Training programs for poll workers have not been a major focus of HAVA grant funding to counties. According to its HAVA spending plan for fiscal year 2006–07, the office distributed $9.9 million in HAVA funds for training programs for poll workers, all in fiscal year 2004–05. However, according to the office’s senior management, a significant portion of the total $365 million HAVA funding was not discretionary and could only be spent on certain activities such as voting systems and a voter information system. Scope and Methodology The Joint Legislative Audit Committee (audit committee) requested that the Bureau of State Audits review the training of poll workers. Specifically, the audit committee asked us to determine the role of the Office of the Secretary of State in providing guidelines or standards to county registrars’ offices, including those that address the training of poll workers. The audit committee also requested that we determine whether these guidelines meet the requirements set in law and regulations, are periodically updated, and address the rights of decline‑to‑state voters. Further, we were asked to determine whether the office has any oversight authority in ensuring that the training guidelines are adhered to. The audit committee also asked us to select a sample of eight counties to identify the methods, format, and timing of the training provided. In addition, we were asked to review the training content and to identify the instructors used for the training and the frequency of the training provided. The audit committee also asked us to determine whether these counties’ training programs complied with the office’s guidance and how often counties update their training. We were also asked to determine whether all poll workers are required to participate in training and whether the counties assess whether poll workers are effectively trained to 12 California State Auditor Report 2008-106 September 2008 deal with any situations that might arise on election day. Finally, the audit committee asked us to determine what actions the counties take when they receive complaints from voters, and how each county determines the number of poll workers to assign to each polling place. To determine the role of the office in establishing guidelines for training poll workers, we reviewed the applicable sections of the Elections Code and interviewed the office’s staff. When considering whether the office’s training guidelines met the requirements set in law and regulations, we evaluated whether the guidelines addressed the eight topic areas required in statute. We also determined whether the office assembled a task force, as required by law, in developing its training guidelines for poll workers. To determine whether the office periodically updates the guidelines, we interviewed the office’s staff. To evaluate the office’s legal authority to enforce a county’s adherence to the training guidelines for poll workers, our legal counsel reviewed applicable laws and discussed this issue with the office’s legal counsel. To evaluate counties’ training of poll workers, we reviewed the practices of a sample of eight counties. The audit committee requested that we include Los Angeles and Kings counties as part of this sample. We judgmentally selected the remaining six counties based on factors such as the number of voter complaints received by the office and the number of registered voters in county jurisdictions. Alameda, Fresno, Orange, San Diego, Santa Clara, and Solano counties were the other six counties we selected for review. To evaluate the counties’ adherence to the office’s training guidelines, we either directly observed, in person or through watching video recordings, poll worker training for the June 2008 election. During these observations, we noted the length of the training and the manner in which the information was presented. Specifically, we considered whether the counties’ training presentations were interactive, based on the use of hands‑on exercises or through questions posed to the audience. To evaluate the content of the training, we developed a checklist to note our observations and ensure that the eight topic areas specified in statute and addressed in the office’s guidelines were covered during training. The checklist is presented in the appendix of the audit report and does not represent an exhaustive list of all of the office’s training guidelines. Instead, the appendix lists those aspects of poll worker training that, in our professional judgment, relate most directly to the audit objectives. Key aspects of poll worker training that we looked for included information pertaining to provisional voting, disabled voters, election challenge procedures, the operation of and troubleshooting of voting equipment, procedures for opening and closing the polling California State Auditor Report 2008-106 13 September 2008 place, voter complaint procedures, and issues pertaining to the prohibition of attempts to influence or intimidate voters when they visit polling places to cast their ballots. Although not specifically addressed in the office’s guidelines, we also evaluated whether counties discussed the voting options of decline‑to‑state voters. Further, to the extent possible, we evaluated whether all poll workers attended training prior to the February 2008 election. To determine how often counties in our sample update their training, and how they evaluate whether poll workers are effectively trained, we interviewed county elections officials and reviewed available documentation, such as training evaluations and postelection debriefing reports, where recommendations for improving poll worker training were discussed. We also identified the sources of information counties use to gain insight into aspects of their poll worker training programs that should be updated, and the extent to which counties could demonstrate that they used the data collected to make such updates. To determine what actions the counties take when they receive complaints from voters, we interviewed county elections officials about their methods and reviewed available documentation. Given that Los Angeles County has the most registered voters in the State, we attempted to review data from its complaint tracking system to evaluate the types of complaints it received during the February 2008 election and the actions the county took to resolve these complaints. The U.S. Government Accountability Office, whose standards we follow, requires us to assess the reliability of computer‑processed data. We attempted to verify the accuracy of the data contained within Los Angeles’ complaint tracking system by selecting a sample of 29 complaints and comparing the data to the county’s corresponding complaint intake forms, which are the data’s source documents and are used by county staff to initially log complaints. Our review found that the action description field for 27 of the 29 complaint forms in our sample did not agree with the source documents. Further, Los Angeles County elections officials acknowledged that data from only about 800 of the approximately 1,000 intake forms used during the February 2008 election were entered into its complaint tracking system. As a result, we concluded that the data in Los Angeles’ complaint tracking system was not sufficiently reliable for the purpose of evaluating how the county responds to voter complaints. Finally, to gain an understanding of how counties determine the number of poll workers to assign to each polling place, we interviewed county elections officials regarding their methodologies. Specifically, we determined what factors counties consider when determining the number of poll workers needed 14 California State Auditor Report 2008-106 September 2008 to work at each polling place during a given election. We also identified whether counties have methods to address potential poll worker staffing issues, such as absenteeism. California State Auditor Report 2008-106 15 September 2008 Chapter 1 The OffICe Of The SeCReTARy Of STATe And OTheR SOuRCeS ASSIST The COunTIeS’ eleCTIOnS OffICIAlS In develOPIng TRAInIng fOR POll WORkeRS Chapter Summary The Office of the Secretary of State (office) has met its statutory obligations to adopt statewide uniform standards for the training of poll workers. Using recommendations from a task force it organized, the office issued its Poll Worker Training Guidelines 2006 (training guidelines) as required by state law. However, the statute that requires the establishment of the training guidelines has no directive for the office to regularly update them. According to the office’s senior management, this is one reason that the office has not revised the guidelines since issuing them in 2006. In addition, the California Elections Code (Elections Code) has no provision requiring the office to monitor the counties’ compliance with the training guidelines. Consequently, poll workers may not receive training that is consistent throughout the State of California (State) as intended by the legislation that requires the training guidelines. The office acknowledges that the training guidelines lack some information related to recent election problems, such as the proper ways to accommodate the rights of citizens who register to vote without designating an affiliation with a political party (decline‑to‑state voters), and it has expressed interest in updating the training guidelines. However, the office anticipates the need to convene a task force similar to the one used to develop the original guidelines, and the parties who will need to participate will not be available until after the November 2008 presidential election. According to the office’s senior management, during the February 2008 presidential primary election, many decline‑to‑state voters were confused about which political parties had authorized them to vote in their primaries. In addition to voter confusion, some news agencies reported that poll workers gave unclear instructions and were unsure as to whether they could volunteer information about which partisan ballots were available to decline‑to‑state voters. Other sources of confusion surrounding decline‑to‑state voters included issues ranging from counties’ inconsistent interpretations of state law to voters who had mistakenly registered for the American Independent Party when they wanted to register as unaffiliated. To avoid further confusion, the office has highlighted the options available to decline‑to‑state voters in its voter information guides. The office has also supported legislation aimed at better informing these voters of their rights. 16 California State Auditor Report 2008-106 September 2008 In addition to providing guidelines for training poll workers, the office offers assistance to county elections officials who develop poll worker training by issuing periodic memorandums (memos) and by providing information during training and informational sessions conducted by other entities. Many of the eight counties we reviewed consider other sources of information for updating their training programs. Seven counties mentioned using the California Association of Clerks and Election Officials (CACEO) as a resource for updating their poll worker training programs. The Election Administration Research Center has also provided tools to counties to develop programs for training poll workers through its surveys of poll workers administered during the 2006 election cycle. The office has also monitored county poll worker trainings and elections through its Election Observation Training Program (observation program), when staff time and funding allowed. The Office of the Secretary of State Lacks a Directive to Update Its Poll Worker Training Guidelines Legislation passed in 2003 required the office to adopt uniform standards for poll worker training. According to the Senate floor analysis of Senate Bill 610 of the 2003–04 regular session of the Legislature, county elections offices were operating independently, without the benefit of established standards and procedures. The legislation called on the office to appoint a task force to study and recommend uniform guidelines for the training of precinct board members and to adopt uniform training standards for precinct board members, also known as poll workers, based upon the recommendations of that task force, by June 30, 2005. In August 2004, the office created the task force, which included the chief elections officers from Los Angeles, Orange, Alpine, Sierra, Yolo, and Contra Costa counties. The task force also included an additional eight members from community‑based organizations that were to have elections expertise and familiarity with different ethnic, cultural, and disabled populations to ensure representation of the State’s diverse electorate. The office provided the Legislature with the task force’s recommendations in March 2005. In February 2006, the office issued a draft of the training guidelines to counties for review, and it issued its final training guidelines on April 19, 2006—almost 10 months late. The Appendix lists the According to the office’s senior eight required training topics, as well as selected training standards management, the office has not from the office’s poll worker training guidelines. revised the training guidelines since issuing the document The office’s senior management stated that the office has not in 2006, pointing to the fact that revised the training guidelines since issuing the document in 2006, the statute includes no directive to pointing to the fact that the statute includes no directive to update update them. them. Our legal counsel agrees with the office’s interpretation that California State Auditor Report 2008-106 17 September 2008 it is not required to update the training guidelines. In addition, the office’s senior management stressed that the office never intended for its guidelines to be comprehensive, because new concerns, such as the issues concerning decline‑to‑state voters that arose during the February 2008 presidential primary, often arise as unique situations that typically cannot be foreseen. The office’s position appears to have merit, since no guideline document can cover every contingency. Further, the CACEO, an organization whose purpose, in part, is to promote high standards and uniformity of practices among local elections officials, believes there is little need to update the office’s 2006 training guidelines because the information that the document contains is still valid. Despite these assertions, senior management with the office have expressed a desire to update the guidelines after the November 2008 election, acknowledging that the office will need to create another task force to do so. The office has nonetheless communicated additional training information to counties through periodic memos to county clerks and registrars of voters. Between April 2006, when the office issued the training guidelines, and April 2008, the office issued more than 650 memos to local elections officials about various election topics. According to the deputy secretary of state for operations, 47 of the memos pertain to poll worker training. However, our review found that only 11 of these memos pertained specifically to poll worker training. Examples of the topics covered in these 11 memos include poll worker etiquette, security provisions for voting equipment, and information that must be posted at polling places. The CACEO indicated that the usefulness of these memos would improve if the office maintained this guidance in a centralized database that elections officials could search by topic area. As of August 2008 the office began posting these memos in chronological order on its Web site. Finally, the office indicated that it has provided additional guidance The office indicated that it has to counties by participating in CACEO‑sponsored conferences, provided additional guidance including the New Law Conference held annually in December, to counties by participating in the CACEO’s annual conference, and the CACEO’s Summer CACEO-sponsored conferences, Institute training course. The office was able to provide the including the New Law Conference, agenda for the trainings it conducted at the New Law Conference held annually in December, in December 2007 which included updates on the federal Help the CACEO’s annual conference, America Vote Act (HAVA), voter registration cards, and office and the CACEO’s Summer Institute outreach activities. However, the office was unable to provide training course. evidence of other trainings it conducted because the presentations were oral in nature, and written documents were not provided to attendees. 18 California State Auditor Report 2008-106 September 2008 The Office Has Taken Steps to Raise Awareness About the Voting Rights of Decline‑to‑State Voters Although the office acknowledged that its poll worker training guidelines do not specifically address the voting rights of decline‑to‑state voters, it has taken steps to guard against the recurrence of the voting problems noted during the February 2008 primary election. Recognizing that decline‑to‑state voters may have been confused over which political party would allow them to vote in the February 2008 primary, the office modified its voter information guide for the June 2008 primary to educate decline‑to‑state voters regarding their rights. Also, two pieces of legislation are currently pending that would require poll workers to inform decline‑to‑state voters of their rights when voting at the polls. Further, as we discuss in Chapter 2, all eight counties we reviewed described the voting options available to decline‑to‑state voters in their training classes prior to the June 2008 election. A decline‑to‑state voter can vote by mail or in person at a polling place. Although a county must notify a decline‑to‑state voter who is on the county’s permanent vote‑by‑mail voter list of their voting options, at a polling place a decline‑to‑state voter will receive a nonpartisan ballot unless he or she requests a ballot of a political party that has authorized such voters to cast ballots in its primary. A voter registered with a specific political party may vote only in that particular party’s primary election. The office’s senior management explained that during the February 2008 presidential primary election, many decline‑to‑state voters were confused about their voting options. Although it is difficult to accurately gauge the extent of the problems that decline‑to‑state voters faced during the February primary election, numerous media outlets reported on the issue. According to the deputy secretary of state for HAVA activities, there are no definitive or statistically valid data that officials can analyze to determine the reason for voter confusion, and all of the data collected by the office were anecdotal in nature. The office’s senior management Although data about the problem are limited, the office’s senior stated that in the February 2008 management stated that in the February 2008 election, many election, many decline-to-state decline‑to‑state voters were confused about their voting options voters were confused about their because only two of California’s six qualified political parties, the voting options because only two of Democratic and American Independent parties, had authorized California’s six qualified political this type of voter to cast ballots in their primaries. According to parties, the Democratic and several media reports, some decline‑to‑state voters did not know American Independent parties, had they could request a Democratic ballot or mistakenly believed they authorized this type of voter to cast could request a Republican ballot, which was not available to ballots in their primaries. them in the February 2008 primary. In addition, the office did not describe the options for decline‑to‑state voters until page 7 of California State Auditor Report 2008-106 19 September 2008 the 32‑page Voter Information Guide. Another source of voter confusion came from voters who believed that they had registered as decline‑to‑state voters but instead had registered with the American Independent Party (AIP). The office’s management stated that many voters have mistakenly believed that belonging to the AIP is the equivalent of being an independent, or decline‑to‑state, voter. In addition to voter confusion, some news agencies reported that poll workers gave unclear instructions and were confused as to whether they could volunteer information about which partisan ballots were available to decline‑to‑state voters. According to the deputy secretary for legislative and constituent affairs, a discrepancy in the Elections Code relating to how decline‑to‑state voters are informed about their right to request partisan ballots added to the confusion. One Elections Code section instructs county elections officials to mail a notice to every decline‑to‑state voter whose name appears on the permanent vote‑by‑mail voter list, informing the voter that he or she may request a ballot by mail for a particular political party if that party adopted a rule authorizing these voters to vote in its primary. The notice also includes a toll‑free phone number voters can call to find out which parties have adopted the open primary rule. However, the state law governing decline‑to‑state voters who choose to cast their ballots at a polling place states that each voter must request a ballot for a party that has authorized decline‑to‑state voters to cast ballots in its primary election. In contrast to the other code section, this section does not direct county elections officials to inform decline‑to‑state voters that they may request a ballot for a particular party. According to the office’s senior management, this inconsistency affected the way decline‑to‑state voters were advised of their ballot options at the polling places, with the advice varying from county to county. The office’s deputy secretary for legislative and constituent affairs reported that during the February 2008 primary, some counties took a proactive approach in notifying decline‑to‑state voters of their options at the polls, while other counties believed that they could not advise decline‑to‑state voters of their ballot choices and could provide a partisan ballot only if the voter requested one. The office has taken steps to eliminate voter and poll worker confusion about the rights of decline‑to‑state voters. The office’s The office’s voter information voter information guide for the June 2008 primary election guide for the June 2008 election emphasized the rights of decline‑to‑state voters. Near the front emphasized the rights of decline- of this document, the office devoted an entire page to clarifying to-state voters, and the office issued the primary process and its relation to the general election, three press releases to educate the defining decline‑to‑state voters as those not affiliated with a electorate on this subject. political party, and disclosing the political parties that have allowed decline‑to‑state voters to vote on their ballots. In addition, 20 California State Auditor Report 2008-106 September 2008 the office issued three press releases between February and June 2008 in an attempt to educate the electorate about the rights of decline‑to‑state voters. Currently, the Legislature is trying to eliminate confusion on the part of poll workers and voters with two pending bills, both of which would require poll workers to inform decline‑to‑state voters of their ballot options. The office supports both bills. The Office of the Secretary of State Provides Limited Monitoring of Training and Elections The Elections Code mandates that the office adopt training guidelines and requires counties to conform their training programs to these guidelines. However, the Elections Code has no provision requiring the office to monitor the counties’ compliance with the training guidelines. The office’s senior management stated that although the secretary of state is the State’s chief elections officer, the law did not establish a monitoring program, direct the office to create a new program, or provide funding or direction to develop regulations that would require adherence to any policies surrounding poll worker training. According to the deputy secretary of state for HAVA activities, while the statute could have been more specific regarding overall oversight responsibilities if such oversight was desired, it appears that the intent of the legislation was for the office to serve as a focal point for bringing counties together to collaboratively develop recommendations for the guidelines. Finally, the office’s senior management stressed that the expertise for the Our legal counsel agrees that the training guidelines came from the counties and that it is ultimately Elections Code does not require the their responsibility to comply with the law. Our legal counsel agrees office to ensure that counties are that the Elections Code does not require the office to ensure that adhering to the training guidelines. counties are adhering to the training guidelines. Despite these views, the office has nonetheless performed some monitoring activities through its Election Day Observation Program (observation program). Created in 2003, the observation program began as a poll‑monitoring program that focused on preventing late opening of the polls, long lines at polling places, and attempts to influence or intimidate voters at the polling places. Subsequent election reviews conducted in 2006 changed the focus to assessing how well counties were meeting the challenges of implementing HAVA requirements. In 2006, the observation program had a goal of providing a more comprehensive snapshot of the election process in selected counties over the course of an election. To help each county identify ways to strengthen its poll worker training and voter education programs, the office shared with each county’s elections officials what observers noted in their respective county. California State Auditor Report 2008-106 21 September 2008 During their review of the February 2008 primary election, observers from the office attended training Observation Program Report Findings for the February 2008 Primary Election sessions for poll workers to see how instructors taught poll workers to deal with issues or problems. The Office of the Secretary of State’s Election Day Observation The office’s senior management stated that they Program noted the following situations related to poll plan to perform another review in November 2008 worker training: but said they are uncertain about the 2010 election • Most poll workers were assigned one set of duties and cycle. According to the deputy director of were trained to accomplish only that set of duties. If operations, whether the observation program will one worker was absent, other workers were unable continue in 2010 is dependent upon available to substitute. resources and whether changes in the law require • Workers lacked clear, step-by-step procedures for changes in polling place operations that dictate a opening, setting up, and closing voting systems. need to observe how the counties are implementing those changes. • Many poll workers did not appear well versed on the security measures for voting systems. The office sent 37 observers to 31 counties during • Some counties told poll workers not to tell the February 2008 election and summarized these decline‑to‑state voters that they could vote either on the observations in an observation report. The report American Independent Party ballot or on the Democratic indicates that the observations were discussed Party ballot. with the participating counties. However, the office • Poll workers received inadequate information about voter missed an opportunity to maximize the value of its privacy issues. findings by deciding to share its results only with • Some poll workers told voters incorrectly that only the 31 counties it reviewed. The CACEO leadership, persons with disabilities could use direct electronic who themselves are county elections officials, stated recording voting machines when, in fact, these machines that they were unaware of the office’s report on are available to any person who asks to vote using the February 2008 election. However, according this method. to the deputy secretary of state for operations, the Source: Office of the Secretary of State’s February 2008 Election office posted it on its Web site in April 2008. Some Day Observation Program Report. of the observation report’s findings are shown in the text box. Surveys of Poll Workers Conducted in 2006 Gave Counties and the Office Another Tool to Improve Training Surveys conducted by the Election Administration Research Center (center) at the University of California, Berkeley, regarding the 2006 primary and general elections provided another resource to counties for improving poll worker training. The office and the center funded the surveys and conducted them in collaboration with the CACEO. The center then issued two reports that offered recommendations on how to improve poll worker training. According to one of the report’s authors, these two reports were issued to the office, the CACEO, and county elections officials. Established in 2005, the center studies elections in California at both the state and local levels and disseminates its findings to elections officials, academic researchers, and the public. The center 22 California State Auditor Report 2008-106 September 2008 conducted a survey of poll workers in California during the June 2006 primary election and again during the November 2006 general election, and it issued two separate reports in September 2006 and May 2007 that included numerous recommendations about how to Some of the Center’s State‑Level improve poll worker training. The center Recommendations Resulting From Its developed its analysis of the June 2006 election by Poll Worker Survey sending surveys to 25 counties, and it received more than 15,408 responses from poll workers. • Consider mandating training for all poll workers. The survey’s results indicated that most poll • Study existing county training and reference materials workers felt that their training adequately for their common characteristics and for well‑developed prepared them for election day, with at least formats, and then develop a boilerplate for basic core text 81 percent responding that they had received and formats that all counties can tailor to their own use. “good” training. Similarly, the center’s analysis of • Encourage counties to set up training programs with the November 2006 election found that at least variations in times, locations, and topics; smaller 74 percent of respondents felt their training classes; and more amenities, such as transportation, was good. refreshments, and comfortable class rooms. • In general, provide more state-level guidance on The center’s report on the November 2006 training curricula. general election also included recommendations to be implemented at the state level, as indicated • Encourage and provide incentives for counties to recruit year‑round, and supply a basic handbook on holding poll in the text box. Some of these recommendations worker orientations for the public. included mandating training for all poll workers and providing more state‑level guidance on Source: Election Assistance Research Center, A Survey of California Poll Workers in the General Election of 2006. training curriculum. According to the deputy secretary of state for operations, many of the recommendations in the report would require direction and funding from the Legislature and the governor before they could be implemented, while others involve changes in policy that would require the cooperation of county elections officials. However, the deputy secretary noted that some of the efforts mentioned in the report have already been undertaken, although perhaps not to the extent envisioned by the center. For example, the deputy secretary of operations stated that the center’s recommendation to develop training curriculum was at least partially addressed by the training guidelines issued in April 2006. Recommendation The Legislature should consider amending the Elections Code to explicitly direct the office to periodically update its poll worker training guidelines and to monitor county adherence to these standards. In the interim, the office should continue with its plans to update its training guidelines and incorporate new guidance on California State Auditor Report 2008-106 23 September 2008 the proper handling of decline‑to‑state voters. Finally, to the extent feasible, the office should continue its efforts to monitor county adherence to its guidelines through its observation program. 24 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 25 September 2008 Chapter 2 AlThOugh geneRAlly COnSISTenT WITh guIdelIneS fOR TRAInIng POll WORkeRS, MOST COunTIeS’ TRAInIng PROgRAMS need IMPROveMenT Chapter Summary The California Elections Code (Elections Code) requires county elections officials to train precinct inspectors concerning their duties for conducting elections, in conformance with uniform training standards (training guidelines) adopted by the Office of the Secretary of State (office). The eight counties we reviewed substantially complied with the content of the office’s training guidelines. However, some counties appeared to only partially train poll workers in certain areas, such as voters’ rights to report illegal or fraudulent activity, sensitivity to diverse cultures, and the operation of voting machines. Additionally, some counties did not use suggested training methods, such as role‑playing for processing voters’ ballots and hands‑on training for learning how to operate voting machines. Although their training programs generally reflect the contents of the office’s training guidelines, three of the eight counties we visited told us they do not believe they are required to follow the training guidelines. One county told us it seldom reviews the training guidelines for current elections because the guidelines have not been updated. The majority of the counties mentioned relying on the California Association of Clerks and Election Officials (CACEO) or the United States Election Assistance Commission (commission) for information with which to update their poll worker training programs. In addition, after encountering problems in the February 2008 primary election with ensuring the rights of voters who registered to vote without declaring a political party affiliation (decline‑to‑state voters), all the counties we visited discussed the available voting options for these voters during poll worker training prior to the June 2008 election. Some of the counties we visited demonstrated noteworthy training practices. Most of these practices seemed targeted toward providing poll workers with additional opportunities to practice what they had learned while also being sensitive to their time commitments. For example, we found that some counties offered training at various times and locations and tailored the content to the experience level of the attendees to promote greater attendance at training. Others offered online training or optional workshops with opportunities for more hands‑on training just prior to the election. 26 California State Auditor Report 2008-106 September 2008 However, not all counties required inspectors to attend in‑person training prior to working during elections while many were unable to provide data demonstrating that they trained all inspectors prior to the February 2008 election. As a result, these counties cannot be certain that their inspectors have the knowledge to efficiently administer election polling places. All of the elections officials at the eight counties we visited indicated that they had some method to collect data to identify needed improvements to their poll worker training programs, and seven were able to provide documentation of some of their methods in practice. Methods the counties use to collect this information include post‑training feedback forms from poll workers, comments from instructors, postelection debriefing reports, analyses of voter complaints, and reviews of questions from poll workers on election day. Although all eight counties asserted that they used some variation of these types of feedback to evaluate their training programs, none could clearly demonstrate how the information collected from the February 2008 election was summarized and led to actual changes in their training programs for the June 2008 election. Moreover, we found that although most of the counties we reviewed discussed procedures for handling voter complaints, the emphasis the counties placed on the information varied. For example, Alameda County used a written handout to inform its poll workers that voters can file complaints with the office or county elections officials, while Orange County devoted time during class to discuss voter complaint procedures. All eight counties also told us they receive complaint calls from voters or poll workers on election day, but only three were able to provide us with a complaint log detailing those calls for either the February 2008 or the June 2008 election. Finally, most of the county elections officials we interviewed used projected voter turnout, estimated absenteeism, and the need for multilingual workers to determine the number of poll workers to assign to each precinct. Many of the counties reported challenges in recruiting an adequate number of poll workers, stating that they relied on practices such as expanding recruiting at schools and hiring a reserve of workers to make up for absenteeism among polling place workers on election day. County Elections Officials Generally Followed the Poll Worker Training Guidelines Issued by the Office of the Secretary of State The training provided by the counties we visited generally complied with the office’s training guidelines. The topics required by the training guidelines are shown in the first column of Table 1. Some California State Auditor Report 2008-106 27 September 2008 of the counties we visited conducted separate and different training programs for inspectors who supervise the polling places and the clerks and students who work with them. Because the Elections Code requires the counties to train only inspectors, the following discussion focuses on this type of training. Table 1 shows whether our sampled counties provided instruction in the eight training topics addressed in the office’s training guidelines, either through classroom instruction or through printed training materials. Table 1 County Compliance With the Office’s Training Guidelines Counties RequiRed tRaining topiCs alameda FResno Kings los angeles oRange san diego santa ClaRa solano Rights of voters Provisional voting Relevant election laws Election challenge procedures Opening and closing the polls Operation of voting systems Cultural competency * Accommodations for disabled voters Source: Auditor observation of county poll worker training sessions for the June 2008 election and review of materials provided during these sessions. Note: Our review did not evaluate county compliance with all the guidelines published in the Secretary of State’s Poll Worker Training Guidelines 2006 document. The Appendix lists the specific guidelines we considered for the training topics shown above. = Trained: poll worker was trained on topic either verbally in class, through the training material provided, or both. = Partially trained: poll worker training was incomplete in a key area on a particular aspect of training. See the Appendix for further explanation and a detailed presentation of the eight training topics. * When evaluating cultural competency we considered whether poll workers were trained to be respectful of diverse cultures and to display multilingual materials. Solano County asserted that it has no foreign language requirements under the Voting Rights Act of 1965. We confirmed the county’s claim by reviewing the list of counties covered under this act as shown in federal regulations. Although all the counties we visited addressed the office’s training guidelines to some extent, Table 1 shows that some counties appeared to only partially train poll workers in certain areas. For example, Fresno County partially trained inspectors on the rights of voters. Specifically, it trained its poll workers on voters’ rights to replace spoiled ballots and their rights to receive assistance when voting, but did not train them on voters’ rights to report illegal or fraudulent activity. In another example, Solano County only partially trained its poll workers on the operation of its voting machines, because it did not discuss how to operate machines intended to assist disabled voters. Although the county’s instructor offered to provide instruction to poll workers on the operation of this machine after class, no one stayed for this training. When asked 28 California State Auditor Report 2008-106 September 2008 why the AutoMARK voting machine was not demonstrated during the training class we observed, Solano County’s instructor stated that “the only reason the demonstration of the AutoMARK was not covered during the class you observed was because we ran out of time. The next best thing we could do was to be available after class for a demonstration, and that offer was made.” Nonetheless, the office training guidelines require instruction in the operation of voting equipment. The Appendix includes a detailed presentation of the eight required training topics included in the office’s guidelines and state law. In addition to generally complying with the office’s content standards, most of the counties we visited complied with the office’s training guidelines regarding the methods and timing of training. The training guidelines state that hands‑on experience is critical to ensure minimal problems on election day. The training guidelines call for workers at polling places that use touch‑screen and optical scan systems to receive hands‑on training on the setup of the system, normal operating procedures, and complete closing procedures. Further, the training guidelines state that training techniques should require poll workers to interact with the trainers and that training should be provided no more than six weeks prior to an election. Table 2 shows the eight counties’ compliance with these requirements. Table 2 County Adherence to the Office’s Training Methods and Timing Suggestions Hands-on questions tRaining no moRe pRaCtiCe WitH VoteR sCenaRio to audienCe to tHan six WeeKs County Voting maCHines Role-playing ReinFoRCe Key points BeFoRe eleCtion Alameda   Fresno   Kings   Los Angeles     Orange     San Diego     Santa Clara     Solano *    Sources: Auditors’ observations of county poll worker training for the June 2008 election and review of materials provided during these trainings. = Adequately adhered to training methods and timing guidelines. * The training session we observed in Solano County provided hands-on practice in using its ballot scanner. However, there was no similar discussion or exercise with its AutoMARK voting system, which the county uses to assist disabled voters. California State Auditor Report 2008-106 29 September 2008 As Table 2 shows, poll worker training conducted by the counties we reviewed generally complied with the training guidelines on methods and timing, with some exceptions. For example, Alameda, Fresno, and Kings counties did not provide role‑playing on how to assist voters in casting their ballots, and Kings, Fresno, and Solano counties did not offer hands‑on practice with all the voting machines during the training, as required by the training guidelines. According to the guidelines, studies have shown that adults learn best in short, interactive training sessions and with hands‑on instruction. In addition, the commission indicated in its July 2007 Successful While a recent training guide Practices for Poll Worker Recruitment, Training and Retention cautions trainers not to expect guide that trainers should not expect that poll workers will read poll workers to read the election the entire election manual provided at training. However, not all training manual, not all of the of the counties we visited verbally covered important topics in their counties we visited verbally poll worker training. For example, Alameda County provided its covered important topics in their instructors with printed material that covered topics not discussed poll worker training. in class. During the training we observed, we noted a lack of discussion pertaining to election challenge procedures, which would inform poll workers of the circumstances under which they could challenge an individual’s right to vote. Further, Alameda County’s instructors did not discuss certain topics pertaining to voters with disabilities, such as a voter’s right to receive assistance from up to two people and the right to vote outside the polling place if the voter cannot leave his or her vehicle. According to Alameda County’s elections officials, training is conducted in this manner because the inspectors’ main responsibilities include opening and closing the polls, as well as managing the security of the vote and the chain of custody documentation. Therefore, the inspectors are not given detailed training in other areas reserved for clerks. However, according to Alameda County it began offering separate training sessions for clerks only for the June 2008 election, although it did not require attendance at these sessions. Similarly, other counties also opted to cover certain election requirements using training handouts as opposed to discussing them in class. For example, Solano County did not verbally instruct inspectors in certain matters specific to disabled voters. Specifically, the instructor did not inform inspectors about their responsibility to make reasonable accommodations for disabled people or verbally inform the inspectors that voters can obtain assistance from up to two other people. In addition, neither Solano County nor Fresno County verbally informed the inspectors that they must allow disabled voters to vote outside the polling place, if requested. Moreover, three of the counties we visited do not consider the training guidelines to be mandatory. Elections officials from Alameda, Solano, and Fresno counties told us they do not 30 California State Auditor Report 2008-106 September 2008 believe they are required to follow the office’s training guidelines. Los Angeles, Santa Clara, Fresno, and Alameda counties stated that they reviewed the training guidelines when the office first established them in 2006, but Fresno County indicated that it seldom reviews them for current elections because the guidelines have not been updated. Instead, many counties in our sample look to other sources of information for updating their training programs. Seven of the eight counties mentioned using the CACEO or the commission for information with which to update their poll worker training programs. All Counties We Reviewed Instructed Poll Workers on the Voting Options of Decline‑to‑State Voters for the June 2008 Election Although not specifically mentioned in the office’s poll worker training guidelines, all eight counties we reviewed discussed the voting options available to decline‑to‑state voters in their training classes prior to the June 2008 election. Following the previous primary election in February 2008, various media reports surfaced detailing voter frustration with poor instructions from poll workers regarding the ballot options available to these voters. In the Election Day Observation Program Office of the Secretary of State’s report of its review of 31 counties during the Observations Regarding Issues Concerning February 2008 election, the office concluded that Decline‑to‑State Voters “many decline‑to‑state voters were confused about (February 2008 election) what voting options were (or were not) available to them, and some poll workers were equally • Many decline-to-state voters were confused about what confused.” Examples of the problems noted by the voting options were (or were not) available to them, and office during its review of the election are shown some poll workers were equally confused. in the text box. The office’s report did not quantify • Some decline-to-state voters asked for American how often these problems occurred, nor did it Independent Party (AIP) ballots, thinking of themselves indicate the counties in which these problems as independents, and not realizing that AIP was a party. were most common. Many voted the AIP ballot and wondered why the candidate they wanted to vote for was not an option. Los Angeles County’s problems with • Some counties told poll workers not to tell decline‑to‑state voting appear to be attributable, in decline‑to‑state voters that they could vote either AIP or part, to the instruction provided to its poll workers Democratic ballots. and the design of its ballot. In its February 2008 Source: Office of the Secretary of State’s Election Day training material, Los Angeles County instructed Observation Program Report for the February 2008 presidential poll workers to tell decline‑to‑state voters, “If primary election. you are voting for candidates in the American Independent or Democratic Party, use the appropriate booth and remember that you must select the party in the first box.” Los Angeles County’s ballot design for decline‑to‑state voters in the February 2008 election has been described by some as a “double‑bubble” ballot, in which decline‑to‑state voters had to indicate both the party they were California State Auditor Report 2008-106 31 September 2008 crossing over to and the candidate they were voting for. According to testimony provided by the Los Angeles County registrar of voters, out of a total of 226,000 nonpartisan ballots cast in the February 2008 election, more than 60,000 were cast for partisan candidates but the voters had not first indicated their intention to cross over to the Democratic Party or the American Independent Party. Through subsequent review, Los Angeles County was able to count slightly more than 48,000 of these 60,000 votes for candidates in either the Democratic or American Independent Party. However, for more than 12,000 votes, the county was unable to identify the intended choice of the voters and include their votes in the election’s final result. Our review of Los Angeles County’s training program for the June 2008 election found that it had taken steps to simplify One improvement that Los Angeles its instructions to poll workers. One improvement that County made was to modify the Los Angeles County made was to modify the design of its ballot design of its ballot by eliminating by eliminating the need for decline‑to‑state voters to first indicate the need for decline-to-state voters which political party’s primary contest they would vote in. Instead, to first indicate which political poll workers in Los Angeles County were trained to provide specific party’s primary contest they would nonpartisan crossover ballots, thus eliminating the need for poll vote in. workers to instruct voters to select both a party and a candidate. The Democratic, Republican, and American Independent parties allowed decline‑to‑state voters to cross over in the June 2008 election. Los Angeles County created a “Nonpartisan‑Crossover Republican” ballot and a “Nonpartisan‑Crossover Democratic” ballot to simplify the process when decline‑to‑state voters wished to participate in these contests. Los Angeles County’s training material clarified that the American Independent Party did not have candidates running for office in Los Angeles County during the June 2008 election, so the county did not create a specific crossover ballot for that party. In addition to changing the design of its ballot, Los Angeles County appears to have placed greater emphasis on explaining the voting options of decline‑to‑state voters during poll worker training. On the first page of the county’s “Election Guide and Checklist” document, a reference document provided to poll workers during training, the county clearly discusses the crossover voting options available to decline‑to‑state voters. The companion document used in the previous February 2008 election did not discuss these issues until page 20 of its 50‑page guide. Similarly, we saw that Los Angeles County appeared to place greater emphasis on decline‑to‑state issues in its training presentations. During the February 2008 poll worker training classes for inspectors, the county devoted one training slide to this issue midway through the presentation. In contrast, the county’s training program for June 2008 devoted multiple slides on the subject in a “What’s New” discussion occurring toward the beginning of the class. 32 California State Auditor Report 2008-106 September 2008 The other seven counties in our sample also addressed decline‑to‑state voting options when providing training to their inspectors. In San Diego County, we noted that instructors provided written examples in class to demonstrate how to check decline‑to‑state voters in on the sign‑in rosters and how to ensure that the correct ballots were provided. Santa Clara County’s training class included a role‑playing exercise it developed to further clarify how decline‑to‑state voters should be assisted on election day. To better inform voters of their voting rights, three counties developed posters to display at each polling place that identified the crossover voting options for decline‑to‑state voters. It is difficult to evaluate whether problems with decline‑to‑state voting occurred in the June 2008 election, since the office had not prepared an observation report for this election at the time we completed our fieldwork in August 2008. However, the Legislature is considering two bills that would require poll workers to inform decline‑to‑state voters of their voting options on election day. Some Counties Exhibited Noteworthy Practices for Training Poll Workers Noteworthy Poll Worker Training Practices Given that counties develop their own training • Various training locations throughout the county. programs, it is not surprising to see variation • A 90-minute refresher training for returning clerks to in how the training is presented. In our review review new information that was not addressed in of eight counties, we observed some notable prior training. practices that other counties might consider. These are listed in the text box. Most of these • Separate training for inspectors and clerks. practices seem targeted toward providing poll • Hands-on and role playing exercises that reinforce workers with additional opportunities to practice key points. what they learned while also being sensitive to • Online training that could be used as an additional their other time commitments. resource to provide information to poll workers and that can be easily accessed. Attending poll worker training can be a substantial commitment of an individual’s free time. Five of • Workshops providing the poll workers an opportunity to practice what they were taught in class and to ask any the eight county poll worker training programs additional questions they might have. we observed lasted three or more hours. In fact, only three of the eight counties had breaks during • Train the trainer programs allowing the counties to make the instruction. In its survey of poll workers sure only the most qualified instructors are used to train during the general election of 2006, the Election poll workers. Administration Research Center (center) at the • Podcasts used as a resource for the county to remind poll University of California, Berkeley, found that the workers of relevant election day information. two primary reasons why poll workers did not Sources: Auditors’ observations of poll worker training for attend training was that they felt it was either the June 2008 election and their review of materials provided unnecessary, since they had already been trained during those trainings. in prior elections, or they had conflicts with the training times offered by the county, such as having to work or being out of town. California State Auditor Report 2008-106 33 September 2008 Some of the eight counties we visited have taken steps to eliminate these barriers by offering different training sessions geared toward experienced and inexperienced poll workers, or by offering various training times and locations. In Orange County, elections officials began offering differentiated training classes for the June 2008 election for new clerks, returning clerks, and inspectors. According to Orange County officials, it took these steps for the June 2008 election based on poll worker feedback from the prior election. The class for returning clerks is 90 minutes long and is meant as a refresher course for clerks who worked during the previous election. In contrast, Orange County’s class for new clerks lasts three to four hours. In addition, offering multiple training locations and starting times is a good practice to help promote greater attendance at training sessions, making it easier for poll workers to fit these trainings into their schedules. In cases where counties may not have the resources to offer multiple training sessions, using online training appears to be an alternative method to provide important information to poll workers. Three counties we reviewed use online training as a supplemental resource. The assistant registrar in Los Angeles County indicated that all poll workers in the county are given access to the online training program, which they can access through the county registrar’s homepage. This training is used to supplement the information that poll workers receive in class. The online program includes various sections such as a library section, a section of frequently asked questions, and a “my training” section. The “my training” section lists various courses, which may contain text, pictures, video, or interactive activities. According to its Web site, most courses also include a brief assessment to help poll workers measure their progress and identify areas needing improvement. The Los Angeles County online program can be accessed as often as poll workers like, and county elections officials encourage its use as a refresher prior to election day. Using online resources in a different way, Orange County offers podcasts, which can be accessed through the county’s Web site. The podcasts inform and remind poll workers of election topics such as information on scanning ballots and other aspects of poll worker training. For some poll workers, their attendance at training may have occurred two or three weeks prior to election day. Given the large amount of information counties are required to provide in these training sessions, it is not surprising that some poll workers may feel unsure of themselves. To help mitigate these concerns and Both Orange and San Diego allow poll workers to feel more confident in their abilities, some counties offered workshops prior to counties offer optional workshops where poll workers could gain the June 2008 election where poll more hands‑on experience and reinforce what was learned in class. workers could gain more hands-on Both Orange and San Diego counties offered these workshops experience and reinforce what was prior to the June 2008 election. According to informational notices learned in class. 34 California State Auditor Report 2008-106 September 2008 provided during training given to San Diego County’s poll workers, these workshops were offered in six locations, and each location was available to poll workers from 9 a.m. until 5 p.m. over the five days preceding the election. By attending these workshops, poll workers could get hands‑on practice in providing the correct ballot to voters, going through various voter scenarios, and setting up the voting equipment. According to the chief deputy of elections services in San Diego County, turnout at these workshops is significant. Finally, some counties we visited hire temporary instructors, who need not have experience with administering elections, to teach their poll worker training classes. In counties that use this approach, it might be reasonable to expect that they would have a process for ensuring that these instructors are sufficiently knowledgeable before training poll workers. Orange and San Diego counties have noteworthy hiring and “train the trainer” processes to ensure that poll worker trainings are staffed with competent instructors. Orange County’s process begins with a formal announcement of the available instructor positions on its Web site. During the interview, prospective instructors are required to demonstrate their teaching ability through a five‑minute presentation. If successful, the instructors begin a three‑week course in which they learn the poll worker curriculum. Finally, prior to teaching poll workers, the instructors must “teach” county personnel and demonstrate that they can successfully deliver the course. San Diego County’s process is similar in that it provides a three‑week training course to trainers and also includes dress rehearsals in front of county personnel. In contrast to Orange and San Diego counties, Alameda County did not have a formal process for hiring its instructors and did not post a job announcement. Nonetheless, according to the county, some individuals learned of the openings and requested the county’s consideration as instructors for poll worker training. However, even though Alameda County hired three instructors, it could not demonstrate that they were adequately trained prior to teaching Although Alameda County poll workers. Although Alameda County claims that its instructors claims that its instructors helped helped construct the poll worker training presentation, our construct the poll worker training observation of one training session found that the instructor was ill presentation, our observation of prepared. On various occasions during the training, the instructor one training session found the was unable to answer questions from the poll worker audience, such instructor was ill prepared. as how many times a voter can spoil a ballot and have it replaced. In other instances, the instructor did not answer audience questions about assisting voters and issuing ballots, stating that such topics would be covered in the training for clerks. However, because state law does not require clerks to attend training, such a response from Alameda County’s instructor was not helpful to the class. To better understand the qualifications of Alameda County’s instructors, we asked the county for the resumes of the three instructors it used for the June 2008 election. Alameda County was able to provide only California State Auditor Report 2008-106 35 September 2008 two resumes, of which only one noted prior experience training poll workers. Ensuring that training is taught by informed instructors seems critical to properly preparing poll workers for election day. Not All Poll Workers Are Required to Attend Training, and Most Counties We Visited Could Not Provide Reliable Training Data Although state law requires that polling place inspectors receive training prior to election day, six of the eight counties we visited were unable to provide reliable data to demonstrate that all of their inspectors had been trained before the February 2008 election. Of the two counties that could provide reliable data, Los Angeles County acknowledged that not all of its inspectors were trained, while Kings County provided evidence showing that all inspectors received training. However, with many of the counties having poor data on training attendance, the extent to which poll workers have the knowledge to effectively administer elections is unclear. In its May 2007 report on poll worker performance during the general election of 2006, the center found that many poll workers who were not trained felt insecure and were overwhelmed by the election process. Further, the center found that poll workers who had to work with untrained poll workers felt that the process was slowed down and that they had to work harder and shoulder too much responsibility. To address these concerns, the center recommended that the State consider mandating training for all poll workers, rather than just inspectors. Many counties we visited were unable to provide reliable data on Many counties we visited were poll worker training, which prevented us from evaluating whether unable to provide reliable data all inspectors were trained prior to the February 2008 election. on poll worker training, which Alameda and San Diego counties provided us with training lists that prevented us from evaluating were incomplete. In Alameda County’s case, we found six instances out whether all inspectors were trained of a sample of 29 in which poll workers had received training but were prior to the February 2008 election. not listed as being trained, while in San Diego County we found only one such instance in our sample. At first glance, these errors may seem harmless, since they indicate that the counties’ records understate the number of poll workers actually trained. However, we found the data to be problematic because using it would potentially result in overstating the number of untrained inspectors who worked on election day. Fresno County did not have training records for the February 2008 election, and its training records for the June 2008 election were also incomplete, with six of the 29 poll workers we sampled having received training but missing from its training list. Orange, Santa Clara, and Solano counties also had difficulty providing us with accurate or consistent training lists for their poll workers. In Orange County’s case, we did not attempt to evaluate 36 California State Auditor Report 2008-106 September 2008 its records, since internal documents from the February 2008 election indicated that its attendance data were inaccurate. Santa Clara County provided two separate training attendance lists for the February 2008 election, both prepared after the election. However, we noted inconsistencies between the two lists, and the county was unable to explain those differences. Solano County was unable to provide us with a list of poll workers who received training prior to the February 2008 election. Instead, the county provided only the sign‑in sheets used at various trainings. Because Solano County did not provide a training attendance list, and we could not verify that it had provided all sign‑in sheets for its poll worker trainings, we did not perform further analysis. Only Los Angeles and Kings counties were able to provide statistics that, based on our audit testing, appeared to be reliable. Los Angeles County’s training data for the February election included both inspectors and clerks, and we could not readily separate the Los Angeles County’s payroll two classes of poll workers in our analysis. Nevertheless, Los Angeles summary report indicates that County’s payroll summary report indicates that 25,631 poll workers 25,631 poll workers worked during worked during the February 2008 election, while the county’s training the February 2008 election, attendance report indicated that 20,610 poll workers—approximately while the county’s training 80 percent—attended the county’s training program. Although these attendance report indicated that statistics may not necessarily indicate that there were untrained 20,610 poll workers—approximately inspectors, Los Angeles County’s assistant registrar acknowledged 80 percent—attended the county’s that “a number of inspectors served when they did not attend training program. in‑person training classes.” Expanding on this comment, the assistant registrar stated, “The causes of this phenomenon are often due to extenuating circumstances such as inspectors resigning on short notice before election day, hard‑to‑recruit areas, and the uniquely large number of inspectors recruited by Los Angeles County.” For the June 2008 election, Los Angeles’ assistant registrar asserted that 3,965 out of 4,282 inspectors were trained, leaving 317 inspectors—or more than 7 percent—who did not attend in‑person training classes. Kings County was able to provide evidence indicating that all 39 of its inspectors received training before the February 2008 election. Counties Collect Data on the Effectiveness of Poll Worker Training From Various Sources In July 2007, the commission issued a report entitled Successful Practices for Poll Worker Recruitment, Training and Retention. In its report, the commission highlighted the practice of obtaining feedback from poll workers as a way to improve training. The commission also recommended reviewing poll worker performance in the last election when planning a training program. By inviting poll workers to provide feedback on the training they received, and surveying them about their experiences on election day, county elections officials would seem better prepared to make meaningful California State Auditor Report 2008-106 37 September 2008 improvements in the training curriculum. County elections officials also have other sources of information they can consider before determining how their training programs for poll workers should be improved. For example, they can consider trends in voter complaints about poll workers, as well as the opinions of county employees who assist with training on the election day process. All eight counties we visited indicated that they had some method of collecting data and identifying needed improvements to their training programs for poll workers, and all but one were able to provide documentation of some of their methods. In Table 3, we list the various direct and indirect methods counties used to collect data on the effectiveness of poll worker training. Direct methods focus explicitly on the effectiveness of the training itself, such as through post‑training feedback forms provided to poll workers and comments from instructors. In contrast, indirect methods focus on postelection analyses that might suggest weaknesses in poll worker training, such as postelection debriefing reports, analyses of voter complaints, or reviews of questions poll workers had when working on election day. As Table 3 illustrates, the eight counties we reviewed varied in their ability to demonstrate that they considered feedback from these different sources. Table 3 Documented Methods Used to Collect Data on the Effectiveness of Poll Worker Training diReCt metHods* indiReCt metHods† papeR-Based eleCtRoniC-Based eValuations oF eValuations posteleCtion summaRized summaRized data suRVeys oF eValuation oF tRaining By tHe oF tRaining deBReiFings data on VoteR on poll WoRKeR poll WoRKeRs online tRaining By instRuCtoRs By County By County Complaints FRom questions/ConCeRns County aBout tRaining poll WoRKeRs teaCHing tHe Class eleCtions staFF eleCtions staFF eleCtion day on eleCtion day Alameda  ‡      Fresno  ‡      Kings  ‡      Los Angeles  §      Orange  ‡      San Diego        Santa Clara        Solano  ‡      Sources: County evaluation documents and assertions.  = Yes, method used. We concluded that a particular county employed the method based on its ability to provide documentary evidence of the method in practice.  = We concluded that a particular county was either unable to provide documentary evidence of the evaluation method or asserted it did not use the method. * Direct methods: methods that focus explicitly on the effectiveness of the training. † Indirect methods: methods that focus on postelection analysis that might suggest weakness in poll worker training. ‡ These counties did not provide online training as a supplemental training resource. § Los Angeles provided evidence of its online evaluations, but indicated that only a minimal number of poll workers were required to submit online evaluations. Los Angeles data showed that approximately 40 poll workers submitted online evaluations. 38 California State Auditor Report 2008-106 September 2008 The most common type of feedback counties considered was direct feedback from poll workers themselves. Six of the eight counties we reviewed were able to demonstrate that they surveyed poll workers about their impressions of the training they attended. The level of detail in these surveys varied among counties. For example, in Alameda County, elections officials asked poll workers whether the material was presented in a manner that was easy to understand, using a 10‑point scale. Alameda County followed up on this question with a few open‑ended questions, such as “What helped you the most to learn the material?” and “What suggestions do you have to improve the presentation?” San Diego and Los Angeles counties asked more questions of their poll workers, such as inquiring as to whether the audiovisual aids were useful and whether the instructors adequately answered questions from the audience. San Diego County’s survey was also unique in that it asked poll workers whether they wanted to attend an additional workshop for more hands‑on experience with voting machines and/or procedures. Poll workers were able to indicate their interest by leaving their contact information. Some counties asked their instructors to comment on how well the training was received. In Orange County’s survey to its instructors, the county asked various questions on topics ranging from the accessibility of the training site to whether any clerks that attended training might be good candidates for an inspector position. The second most common type The second most common type of data collected, after poll worker of data collected, after poll surveys, were postelection debriefings as an indirect means of worker surveys, were postelection identifying opportunities to improve poll worker training. Five of debriefings as an indirect means the eight counties we visited were able to provide documentation of identifying opportunities to of their postelection debriefings. In all five cases, we saw that the improve poll worker training. counties discussed either positive aspects of training or areas that needed further refinement. San Diego County’s review of its performance during the February 2008 election considered both. Specifically, San Diego County identified “what worked well” and “what we can improve” for multiple aspects of its training program, including its “train the trainer” program and instructor staffing issues; training locations; class scheduling; training content; and online training program. Table 3 also indicates that relatively few counties in our sample could demonstrate that they collected feedback from voters or poll workers on election day. Only Los Angeles and San Diego counties were able to demonstrate that they obtained and summarized data regarding voter complaints. Further, only Los Angeles, Orange, and Alameda counties were able to demonstrate that they did the same for questions or concerns from poll workers during the election. By analyzing data from these sources, these counties could potentially identify aspects of their training programs for poll workers that need refinement. However, while collecting and analyzing data California State Auditor Report 2008-106 39 September 2008 on the effectiveness of poll worker training is a good practice, the value of these data is diminished when they are not used to make improvements. None of the Eight Counties Could Demonstrate How They Identified Changes Needed in Poll Worker Training All eight counties in our sample either asserted that they collected various types of data to evaluate their training programs for poll workers or provided examples of the feedback collected, such as postelection evaluation forms from poll workers and county observations of polling places. However, Alameda, Kings, Los Angeles, and Solano counties could not provide documentation of any analysis of the February 2008 election that would identify the need for changes to their training of poll workers for the Fresno, Orange, San Diego, June 2008 election. Fresno, Orange, San Diego, and Santa Clara and Santa Clara counties were counties were able to provide postelection evaluation reports that able to provide postelection described what changes were needed in their training programs for evaluation reports related to poll workers; however, these reports did not link the conclusions the February 2008 election that based on the data collected with the proposed changes to be made. described what changes were As a result, we could not determine whether the counties in our needed in their training programs sample effectively used the data they collected to improve poll for poll workers; however, these worker training, or whether they are missing opportunities to use reports did not link the conclusions additional data from other sources. based on the data collected with the proposed changes to be made. The eight counties in our sample varied in their ability to provide documentation that explained how they update their poll worker training. Alameda, Los Angeles, Kings, and Solano counties all indicated that their process for updating poll worker training was not always documented or was based on the “institutional knowledge” of county staff, or a combination of these factors. In Alameda County, elections officials asserted that they issue a survey to poll workers following an election to identify areas in need of improvement. Alameda County also asserted that it conducts postelection meetings with county staff to identify any observations. However, the county could not locate its survey or its summarized results, acknowledging that its postelection debriefings are informal. Los Angeles County indicated that it holds numerous debriefings on a continual basis, but that these discussions are not always documented. Specifically, the assistant registrar indicated that “in past major elections all information from surveys, Election Day assessments, attendance of training, etc. are compiled to create a report [on] the effectiveness of Election Day . . . The report for the February [2008] election had not been compiled because of the timing of the next election and [it] wouldn’t be completed until after the June election.” Los Angeles County’s assistant registrar also 40 California State Auditor Report 2008-106 September 2008 clarified “that election review and election review meetings are a daily and ongoing process but that election review documentation is not a . . . formal process given that it is constantly ‘in process’ and that ‘in process improvements’ that become formalized may be significantly misconstrued as final.” Los Angeles County was able to provide us with an example of a postelection evaluation report from the November 2006 election. In this report, Los Angeles County evaluated precinct performance in certain areas, such as whether voting booths were set up correctly, provisional voting lists were used properly, and ballot statements were completed correctly. In some cases, the sources for this information were county personnel, referred to as coordinators, who roamed between different polling places on election day. In other cases, data came from the county’s review of the election material and ballots that were turned in after Although the data in Los Angeles the election. Although the data in Los Angeles County’s report County’s report suggested areas suggested areas where poll worker training could potentially be where poll worker training could improved, it is unclear from this document what specific areas potentially be improved, it is unclear of training the county intended to change. We noted similar from this document what specific circumstances in Kings and Solano counties. Kings County believes areas of training the county intended that its informal, undocumented method is adequate because the to change. county is relatively small, with few poll workers. Fresno, Orange, San Diego, and Santa Clara counties were all able to provide some kind of postelection debriefing or evaluation reports from the February 2008 election that documented areas for improvement in subsequent poll worker trainings. Further, in August 2008 Los Angeles County was ultimately able to produce a postelection debriefing report for the February 2008 election. Orange County’s “Lessons Learned from February Election” document listed several recommended changes based on surveys of instructors and county staff. However, the tabulated results from these surveys were not mentioned in the document. Suggested changes were general in nature and included items such as separating classes for returning and new poll workers and making training a more hands‑on experience. However, it did not appear that Orange County’s document considered feedback from voters, such as complaints, which may have highlighted areas where poll workers were not adequately informed or misunderstood election procedures. San Diego County’s debriefing document was very similar in form to Orange County’s analysis, containing suggestions on how training could be improved. San Diego County’s chief deputy registrar asserted that the county evaluates the effectiveness of its poll worker training by reviewing evaluations from poll workers, online training results, evaluations of trainers by county staff, voter complaints, and other means. However, based on the documents California State Auditor Report 2008-106 41 September 2008 San Diego County provided, it was not clear whether these sources were used and how the information from these sources resulted in actionable decisions to improve poll worker training. Specifically, we did not see data containing comments from voters, poll workers, and county staff discussed or identified as a source of information in its debriefing document. As a result, it is unclear how data from these sources influenced San Diego County’s ideas for training improvements. Areas where San Diego County noted the potential for improvement included making training more fun with games and covering information on common mistakes and how to prevent them. Similar to those of Orange and San Diego counties, Santa Clara County’s debriefing document had bulleted suggestions for improvement, such as “need more training on procedures— role‑playing” and “more [polling place] closing instruction.” In Fresno County, the assistant registrar indicated that “the county holds a wrap‑up meeting with all of the county staff who were involved with any aspect of the election. Each area of the election process is discussed, and all staff have input in the meeting.” However, when we asked Fresno County to provide any documents describing the outcomes of these meetings and their impact on Fresno County provided a poll worker training, Fresno County indicated that it does not have debriefing document describing “specific documents available.” Instead, Fresno County provided a what aspects of training needed debriefing document describing what aspects of training needed improvement, but it is unclear improvement, but it is unclear what data were used to reach what data were used to reach these conclusions. these conclusions. When possible, we tried to identify instances in which counties made changes to their training programs by comparing training materials from the February and June 2008 elections. With the exception of Los Angeles County, which emphasized the proper way to assist decline‑to‑state voters in its June 2008 training guide, and Orange County, which offered different training classes for new and experienced poll workers, it was difficult to identify areas where counties made substantive changes. For example, we could not identify any significant change in Fresno and Santa Clara counties’ training programs. In Alameda County, elections officials indicated that the county needed to revamp its entire training after the February 2008 election, but it could not provide any specific reasons for its revisions. When we examined the materials to see what changes Alameda County made, we noted that it increased the length of its guide from 25 pages to 61 pages, adding a section devoted to the responsibilities of inspectors and expanding its existing guide on troubleshooting voting machines. 42 California State Auditor Report 2008-106 September 2008 Many Counties Did Not Keep Data on How They Responded to Complaints Under state law, voters have the right to ask poll workers and elections officials questions about election procedures and to receive an answer or be directed to an appropriate elections official for an answer. As a condition to receiving federal HAVA funds, federal law also requires the secretary of state to have a plan to deal with voter complaints, which California satisfies by having a toll‑free number that voters can call on election day. This toll‑free number is provided on a “Voters Bill of Rights” notice that must be posted in all polling places across the State. As noted in the Appendix, our review of the training practices at eight counties found that most discussed procedures for handling voter complaints. However, the emphasis that counties placed on this information varied. For example, Alameda County did not discuss complaint procedures during the training class we observed, choosing instead to inform poll workers on pages 10 and 11 of its 61‑page training guide that voters have the right to call the secretary of state or the county elections office. In contrast, Orange County devoted time during its training presentation to discussing voter complaint procedures and developed voter assistance referral cards to be handed out to voters who complained at the polls. These cards provided the telephone number of a voter hotline staffed by the county’s elections staff. We noted that other counties, such as Los Angeles, San Diego, and Santa Clara, developed similar reference cards. Although all eight counties we Although all eight counties we visited indicated that they can visited indicated that they can receive complaint calls from voters or poll workers on election day, receive complaint calls from voters not all take steps to document the complaints that come in and or poll workers on election day, how these complaints are ultimately resolved. Only three of the not all take steps to document eight counties we visited were able to provide us with a complaint the complaints that come in log detailing calls from poll workers, voters, or both during the and how these complaints are February or June 2008 election. Alameda County’s complaint ultimately resolved. data only indicated what the complaints were about, such as disruptive poll workers or concerns about voter registration. San Diego County provided us with examples of complaint logs detailing how it responded to voter and poll worker complaints. In some cases, San Diego County’s records indicated that its voter complaints were resolved by phone, email or other means. In other cases, information about how complaints were resolved was missing. Los Angeles County was also able to provide us with information from its complaint database. Since Los Angeles County has the most registered voters when compared to the other seven counties in our sample, we attempted to further analyze data from Los Angeles County’s complaint database to assess how it responded to complaints. However, given our concerns with the California State Auditor Report 2008-106 43 September 2008 reliability of the data in Los Angeles County’s complaint database, as we discuss below, and that six counties in our sample could not provide information on how complaints were resolved, we could not evaluate the steps these counties took to resolve voter and poll worker concerns. Some county officials saw limited value in recording incoming complaint information and how it was resolved. According to the registrars of Kings and Fresno counties, the number of complaints received is so small that they can rely on their election teams’ collective memory to best utilize the information and appropriately address voters’ concerns. Although Santa Clara County was able to provide some records of the complaints it received, the county’s election services coordinator told us that summarizing complaints would be labor intensive and would take time away from the county’s existing resources. Los Angeles County tracks complaints received during an election using a priority‑based system to collect and resolve voter and poll worker complaints. As calls came in during the February 2008 election, the details of each call were recorded on color‑coded forms to denote the priority level of the call. For example, high‑priority calls would include issues such as a power failure at a polling place, an inspector who never arrived for duty, or broken voting equipment. The second priority level includes calls from poll workers who are confused about procedures or calls from voters who are concerned about their ability to vote. Regardless of the priority level, the calls are logged on the front of the colored form while the action taken and the elapsed time to resolve the issue are recorded on the back of the form. After the February 2008 election, Los Angeles County elections staff entered the information from these forms into a complaint database. Los Angeles County’s ability to use the information in its complaint Los Angeles County’s ability to use database to determine how effectively and promptly it resolved the information in its complaint complaints is limited, however, primarily because its data are database to determine how inaccurate and incomplete. Out of a sample of 29 complaints from effectively and promptly it resolved the complaints database, we noted that the action taken to resolve the complaints is limited, primarily complaint was missing on 27 priority forms. However, for 18 of because its data are inaccurate these 27 priority forms, Los Angeles County’s complaint database and incomplete. indicated that the complaint was “solved by phone.” When we asked Los Angeles County officials to explain this discrepancy, the assistant registrar explained that this entry in the database was the default used when a more detailed explanation of a resolution was not available. We also discovered that Los Angeles County did not enter data from all of its complaint forms into its database. According to the assistant registrar, after the election, the employee entering the forms into the database does not enter ones that appear to be duplicates or dropped calls. However, the assistant registrar indicated that these forms are not kept. Therefore, we could not be certain that all substantive calls were recorded in the county’s complaint database. 44 California State Auditor Report 2008-106 September 2008 Further, the assistant registrar acknowledged that information from only about 800 of the approximately 1,000 complaint forms from the February 2008 election were entered into its complaint database. According to the assistant registrar, a new electronic complaint processing system and an election‑day troubleshooting system that does not use paper forms was used for the June 2008 election. The assistant registrar stated that this system—in which staff enter information into an online database—is more efficient and reliable in terms of its ability to help ensure that multiple resources are not being deployed for the same problems, to empower staff to solve problems using online resources, and to enable staff to more quickly escalate complex problems to experts. Counties We Visited Staff Polling Locations With Poll Workers Based on a Variety of Factors Under the Elections Code, county elections officials are responsible for defining the boundaries of election precincts and establishing precinct boards consisting of the poll workers for a given precinct. At a minimum, one inspector and two clerks must be assigned to every precinct board. County elections officials may assign additional poll workers in proportion to the number of registered voters within a precinct, but no precinct may have more than 1,000 voters. While these requirements seem straightforward, in practice county staffing decisions are more complex. Under state law, counties must also The counties we visited cited consider whether at least 3 percent of the registered voters served various factors when considering by a precinct board speak a particular language other than English. how many poll workers to recruit If so, counties are required to make reasonable efforts to recruit poll for election day, such as projected workers who are fluent in both English and these other languages. voter turnout, expected poll worker The counties we visited cited various factors when considering how absenteeism, past experience many poll workers to recruit for election day, such as projected voter with elections, and the need for turnout, expected poll worker absenteeism, past experience with multilingual poll workers. elections, and the need for multilingual poll workers. When asked to describe their methodologies for determining the number of poll workers required for an election, all eight counties described general poll worker recruitment goals such as ensuring that each precinct has at least one inspector and three to four clerks. Los Angeles County provided us with a slightly more detailed approach that establishes poll worker recruitment targets based on the size of a precinct. Under Los Angeles County’s model, a precinct with 500 or fewer voters will be assigned four poll workers. Precincts with more than 500 but fewer than 700 voters will be assigned five poll workers. Finally, precincts with more than 700 voters will be assigned six poll workers. Table 4 illustrates the various recruitment goals for poll workers at the eight counties we visited and the counties’ stated methodologies for how they arrived at those goals. California State Auditor Report 2008-106 45 September 2008 4 elbaT noitcelE 8002 yraurbeF eht rof seigolodohteM dna slaoG tnemtiurceR ytnuoC fo elpmaS detRopeR-ytnuoC llop egaReVa Fo ReBmun Rep sReKRoW llop egaReVa sReKRoW llop detRopeR-ytnuoC detRopeR-ytnuoC liam-yB-etoV-non Rep sReKRoW desu Ro detnioppa liam-yB-etoV-non Fo ReBmun ReKRoW llop *tCniCeRp *eCalp gnillop yad noitCele no *stCniCeRp *seCalp gnillop laog tnemtiuRCeR laog tnemtiuRCeR eViRed ot desu ygolodoHtem ytnuoC llop ruof htiw ecalp gnillop hcae ffats ot deirt dah ti taht detressa ytnuoC ademalA ademalA rof deen ehT .)rotarepo enihcam gnitov eno dna ,skrelc owt ,rotcepsni eno( srekrow dna tuonrut retov detcepxe no desab saw etis ralucitrap a ta srekrow llop lanoitidda 9.4 9.4 649,3 508 508 000,4 .ecneirepxe tsap srekrow llop ruof htiw ecalp gnillop hcae ffats ot deirt ti taht detressa ytnuoC onserF onserF 2.4 0.6 638,1 934 703 657,1 .)skrelc eerht dna rotcepsni eno( erusne ot seirt dna secalp gnillop 93 ro 83 sah yllausu ti taht detressa ytnuoC sgniK sgniK .ecalp gnillop hcae ot dengissa srekrow llop xis dna ruof neewteb eb lliw ereht taht gnillop hcae ta srekrow llop eht fo eno ,reganam snoitcele s’ytnuoc eht ot gnidroccA llop laugnilib eno tsael ta tiurcer ot seirt osla ytnuoc ehT .rotcepsni na eb ot si ecalp 6.4 7.4 871 93 83 651 .ecalp gnillop hcae ta sretov gnikaeps-hsinapS tsissa ot rekrow desab srekrow llop fo srebmun defiiceps stegrat ti taht detacidni ytnuoC selegnA soL selegnA soL ,sretov deretsiger 005 naht rewef sah tcnicerp a fI .tcnicerp a ni sretov fo rebmun eht no stiurcer ti ,sretov 996 dna 005 neewteb sah tcnicerp a fI .srekrow llop ruof stiurcer ti llop xis stiurcer ti ,sretov deretsiger erom ro 007 sah tcnicerp a fI .srekrow llop evfi ot selegnA soL esuac yam stnemeriuqer laugnilib dna snoitatcepxe tuonruT .srekrow 9.5 9.5 969,52 973,4 973,4 338,42 .tcnicerp nevig a ta stegrat eseht naht erom tiurcer rep srekrow llop ruof fo muminim a seriuqer ti taht detacidni ytnuoC egnarO egnarO 4.3 5.5 561,6 738,1 321,1 000,6 .etis gnillop hcaE .tcnicerp hcae ot srekrow llop ruof sngissa ti taht detroper ytnuoC ogeiD naS ogeiD naS .srekrow lanoitidda eerht dna rotcepsni tcnicerp a ,muminim a ta ,evah lliw tcnicerp fo rebmun eht no desab srekrow llop laugnilib gningissa sredisnoc osla ogeiD naS 1.4 1.5 648,6 056,1 633,1 326,6 .sretov gnikaeps hsilgnE-non hcae ta skrelc eerht dna rotcepsni eno secalp ti taht detats ytnuoC aralC atnaS aralC atnaS 3.6 9.6 400,4 046 185 487,2 .elbissop fi skrelc ruof evah ot sreferp dna ecalp gnillop srekrow llop evfi fo smaet ecalp ot stpmetta ti taht detressa ytnuoC onaloS onaloS no decalp si sisahpme ralucitraP .tcnicerp hcae ta )skrelc ruof dna rotcepsni eno( 3.4 8.5 495 831 201 655 .ecalp gnillop a gnirahs stcnicerp rof krelc htruof a gnitiurcer .sretov fo srartsiger ytnuoc eht morf deniatbo snoitressA :secruoS .slaicffio snoitcele ytnuoc yb dedivorp snoitressa no desab serugfi detiduanu era stnuoma evoba ehT :etoN ereht nehW .ecalp gnillop a ot tcnicerp eno naht erom ngissa yam slaicffio snoitcele ytnuoC .noitcele hcae ot roirp slaicffio snoitcele ytnuoc yb denimreted seiradnuob lacihpargoeg yb denfied era stcnicerP * .dengissa si ecalp gnillop on esac hcihw ni ,tcnicerp liam-yb-etov a ti etangised yam slaicffio ytnuoc ,tcnicerp a ni sretov deretsiger ssel ro 052 era 46 California State Auditor Report 2008-106 September 2008 According to the registrar of voters in Kings County, one of the challenges they face is ensuring that they recruit enough poll workers, since workers may unexpectedly decide not to show up on election day. This problem does not appear to be localized to this county. In other counties, we noted that elections officials also take steps to address this issue. For example, some counties recruit poll workers as reservists who can be deployed to understaffed polling places. According to the community outreach manager in Orange County’s registrar of voters, the county sets a minimum goal of recruiting 100 reservist poll workers who go through poll worker inspector training. Individuals of this group, called the “A‑team,” can be deployed as needed from the county’s election‑day command center. The assistant registrar for Los Angeles County indicated that they have a similar practice, maintaining a reservist group of 400 poll workers. These reservists are sent out to polling places that do not meet the minimum of three poll workers or when a polling place indicates that it needs extra help. Our discussions with Alameda and San Diego counties indicated that they too try to recruit reservist poll workers. Although six of the counties in our sample try to recruit groups of reservist poll workers, and all eight counties try to recruit more than the minimum of three poll workers per precinct, one county in our sample intentionally recruits more poll workers than it needs, to compensate for poll workers who do not show Staff from Santa Clara’s registrar up for duty. Staff from Santa Clara’s registrar of voters indicated of voters indicated that they that they recruit 5 percent more inspectors and 10 percent recruit 5 percent more inspectors more clerks than are called for based on their methodology of and 10 percent more clerks assigning one inspector and three poll workers to each precinct. than called for based on their Indicating that it had 640 precincts during the February 2008 methodology of assigning election, Santa Clara County explained that its recruiting goals one inspector and three poll were 672 inspectors and 2,112 clerks, for a total recruitment workers to each precinct. goal of 2,784 poll workers. Other counties’ poll worker recruitment strategies included seeking out qualified students to work on election day. The Elections Code allows county elections officials to appoint no more than five students per precinct when they meet certain requirements, such as being at least 16 years old, being a U.S. citizen, and having a grade point average of at least 2.5. Five counties stated that they used this option. For example, elections officials from Fresno and Orange counties indicated that they enlisted students to help address poll worker staffing concerns. Fresno County distributes information pamphlets to students, encouraging them to sign up as student poll workers by telling them that they may earn community service credits toward graduation and that the experience looks great on job and college applications. Orange County has a student recruitment initiative called MyBallot to introduce students to the procedures associated with conducting an election. This program California State Auditor Report 2008-106 47 September 2008 is intended to encourage students to participate in training and to work as poll workers on election day. According to its Web site, the MyBallot program builds on students’ skills in English, math, and civics while offering insights into the complexity of elections. Through the program, Orange County reported that it was able to recruit approximately 500 student poll workers for the June 2008 primary election. Although Table 4 on page 45 indicates that most counties reported being able to meet their recruiting targets for the February 2008 election, Los Angeles County was still trying to meet its goals only days before the election. The county’s registrar‑recorder/county clerk issued a press release in January 2008, asking for more poll worker volunteers. The press release stated that although it had recruited more than 24,000 poll workers, there was a need for more. In fact, based on its goal of having 25,000 poll workers, Los Angeles County appears to have been short 1,000 poll workers just days before the election. According to the assistant registrar in Los Angeles, returning poll workers in Los Angeles County make up about 60 to 80 percent of all poll workers working in each election. Based on Los Angeles’ goal of having 25,000 poll workers for the February 2008 election, that equates to having to recruit between 5,000 and 10,000 new poll workers. Recruiting challenges are not limited to Los Angeles County. Six days before the June 2, 2008, election, Fresno County issued a press release indicating that it was 150 poll workers short of its goal. When we spoke with Fresno elections officials, they stated that they had a goal of recruiting approximately 1,700 poll workers for the June 2008 election; however, more than 1,000 people who initially signed up to be poll workers dropped out. Fresno County told us that the dropout rate is usually 500 to 600 poll workers. With concerns over poll worker absenteeism, the recruitment of poll workers is a continual process for some counties, with staff dedicated to the effort. For example, Los Angeles County employs 115 full‑ and part‑time recruiters whose responsibility it is to recruit poll workers throughout the year. Recommendations To ensure that poll worker training programs conform with the office’s guidelines, county elections officials should review the content of their programs, ensuring that their training fully covers topics such as voter complaint procedures, preventing voter intimidation, and issues pertaining to a culturally diverse electorate. 48 California State Auditor Report 2008-106 September 2008 To improve poll workers’ willingness to attend training and their ability to retain the lessons learned, county elections officials should consider implementing the following practices: • Maximize the number of training sessions scheduled for poll workers while also offering the training at multiple locations with different start times to better accommodate poll workers’ other time commitments. Also, providing condensed training tailored to experienced poll workers may entice greater attendance, while more extensive training can be reserved for new poll workers. • Offer poll workers an opportunity to reinforce what they learned in class through the use of online supplemental training material. Such an online program might include practice quizzes on election‑day procedures, examples of the election materials to be used, and reference materials provided at training. County elections officials might also consider providing podcasts that emphasize critical aspects of poll worker training. • Provide optional workshops giving poll workers additional opportunities to practice what they learned and to get hands‑on experience in the use of election‑day supplies and voting equipment. County elections officials might consider providing these workshops 0n the days immediately before an election to maximize poll worker confidence and retention of information. • Take steps to ensure that instructors are sufficiently knowledgeable about county election procedures prior to allowing them to teach the class. For example, county elections officials might consider first requiring their instructors to practice teaching the material to county personnel before they actually train poll workers. To better ensure that county elections officials provide knowledgeable inspectors to serve voters, counties should take steps to ensure that all inspectors receive training. Steps that counties might take to achieve this goal include: • Compiling accurate lists of inspectors who attended training, while informing inspectors who did not go through training that they cannot serve as inspectors. • Recruiting reserve poll workers who have gone through inspector training to be deployed, as necessary, to polling places where the assigned inspectors did not receive the required training. To better ensure that training programs for poll workers are effectively evaluated and needed improvements identified, county elections officials should consider taking steps to track voter California State Auditor Report 2008-106 49 September 2008 complaints and poll worker questions that are received during an election, evaluate whether such comments suggest ways to improve their training programs, and implement those improvements. We conducted this review under the authority vested in the California State Auditor by Section 8543 et seq. of the California Government Code and according to generally accepted government auditing standards. We limited our review to those areas specified in the audit scope section of the report. Respectfully submitted, ELAINE M. HOWLE, CPA State Auditor Date: September 18, 2008 Staff: Grant Parks, MBA, Project Manager Norm Calloway, CPA Rosa I. Reyes Aaron Fellner, MPP Melissa Arzaga Roye, MPP Linda M. Lavin, MPP Mark Needham Wesley Opp Maya Wallace For questions regarding the contents of this report, please contact Margarita Fernández, Chief of Public Affairs, at 916.445.0255. 50 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 51 September 2008 Appendix COunTIeS geneRAlly COMPly WITh TRAInIng guIdelIneS fROM The OffICe Of The SeCReTARy Of STATe RelATIng TO The SubjeCTS COveRed In The InSPeCTOR TRAInIng ClASS State law requires that uniform standards for training poll workers include the following topics: rights of voters, provisional voting, relevant election laws, election challenge procedures, opening and closing the polls, operation of voting systems, cultural competency, and accommodations for disabled voters. The Office of the Secretary of State (office) included these topics in its poll worker training guidelines (training guidelines), which it issued in 2006. Through observing selected training sessions for poll workers, and by reviewing election training materials for June 2008, we assessed whether the eight counties we visited complied with certain aspects of the office’s training guidelines. For each of the eight topics, we focused only on selected areas, as discussed in the Scope and Methodology section of the report. Table A on the following page provides an overview of the counties’ compliance with aspects of the office’s training guidelines, based on our review of how the counties addressed each topic in class or in the training materials that they provided to poll workers. 52 California State Auditor Report 2008-106 September 2008 Table A Selected Counties’ Compliance With the Office of the Secretary of State’s Training Guidelines Counties RequiRed tRaining topiCs and tHe seleCted guidelines ReVieWed* alameda FResno Kings los angeles oRange san diego santa ClaRa solano Rights of Voters Poll workers trained on voters’ right to replace spoiled ballots         (up to 3 times) Poll workers trained on voters’ right to report illegal/fraudulent         activity (poll workers trained on how to provide voters this option) Poll workers trained on voters’ right to have assistance while voting         (up to two people) Provisional Voting Poll workers trained on a voter’s right to cast a provisional ballot         Poll workers trained on when provisional voting is required         Relevant Election Laws Poll workers trained on election-day duties         Poll workers trained to prohibit electioneering within 100 feet of a         polling place Poll workers trained on prohibiting the intimidation of voters at the polls         Election Challenge Procedures Poll workers trained on when and how they can challenge an         individual’s right to vote Opening and Closing the Polls Poll workers trained on setup of the polling place and when to open         the polls Poll workers trained on closing the polls and how to handle lines of         voters at closing Operation of Voting Systems Poll workers instructed on operation of voting machines (so as to        † assist/educate voters) Poll workers trained on troubleshooting voting machines         Poll workers trained on setup and disassembly of voting machines         Poll workers trained on how to recognize/detect tampering with         voting machines Cultural Competency Poll workers trained to be polite to voters and respectful of   ‡      diverse cultures Poll workers trained to display multilingual materials        § Accommodations for Disabled Voters Poll workers trained on the option of curbside voting for the disabled         Poll workers trained on the operation of voting systems intended to         assist disabled voters Source: Auditor observation of county poll worker training sessions for the June 2008 election and review of materials provided during these training sessions.  = Poll worker was trained on topic (either verbally in class, through the training material provided, or both). We indicated when we judged that training lectures or documents provided to poll workers adequately covered these areas.  = Neither verbal instruction nor written training materials covered this particular aspect of training. * These categories are not an exhaustive list of requirements published in the Office of the Secretary of State’s Poll Worker Training Guidelines 2006 document. Our review evaluated only whether poll worker training programs addressed the selected guidelines shown. † Solano County provided hands-on training on how to operate the ballot scanner. However, it did not offer training on how to operate its AutoMARK voting machine. ‡ Kings County offered a slide in its PowerPoint presentation entitled “Respect” though it does not specifically address strategies on dealing with diverse cultures. § Solano County asserted that it has no foreign language requirements under the Voting Rights Act of 1965. We confirmed the county’s claim by reviewing the list of counties covered under this act as shown in federal regulations. California State Auditor Report 2008-106 53 September 2008 (Agency response provided as text only.) Secretary of State State of California 1500 11th Street, 6th Floor Sacramento, CA 95814 September 5, 2008 Ms. Elaine M. Howle, State Auditor Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: Thank you for the opportunity to comment on your recent audit report regarding county poll workers. Overall, I concur with the recommendations made by the Bureau of State Audits as they relate to the operation of the Secretary of State’s office. As I was not permitted to review the portions related to your staff’s audit of county elections offices, I do not have any comments on those sections of the audit report. I would like to make the following comments in an effort to clarify some provisions of your audit report. (1) The report mentions in several places that the state Legislature has sent to the Governor two measures requiring poll workers to notify decline‑to‑state (sometimes referred to as nonpartisan) voters of their ability to receive a ballot for any political party that allows decline‑to‑state voters to vote in their primary election. I would note that I sponsored one of those two measures, specifically writing the legislation and asking an Assemblymember to introduce it on February 22, 2008, less than three weeks after many decline‑to‑state voters encountered problems at some polling places in California. (2) The report notes that state law only requires polling place inspectors, and not all polling place workers, to receive training in order to work at the polls on Election Day. This is accurate, but I believe it is helpful to provide some context to this issue. There were 23,109 precincts for the February 5, 2008, statewide election, but because some polling places house more than one precinct and other precincts only permit voters to vote by mail, there are about 20,000 polling places that must be staffed by nearly 100,000 poll workers at each statewide election. Most counties have an average of three or four poll workers for every one polling place inspector, meaning the vast majority of people working at the polls are not legally required to attend training sessions. All county elections officials conduct training sessions and encourage all poll workers to attend those sessions. Some county elections officials do require all poll workers to attend those sessions and I am certain that all county elections officials would love to require training for all poll workers. Many county elections officials though are faced with a shortage of poll workers and such a requirement could create further staffing shortfalls. (3) The report notes that data collected by the Secretary of State’s office regarding the confusion and difficulties many decline‑to‑state voters encountered at the polls in February 2008 was “anecdotal in nature” and that there was no “definitive or statistically valid data” that could be analyzed to determine the reasons for voter and poll worker confusion. 54 California State Auditor Report 2008-106 September 2008 Ms. Elaine Howle September 5, 2008 Page 2 It is true that the Secretary of State’s office has no way to determine if the complaints were anomalies, or representative of problems in an entire county or all of the state’s 58 counties. I only have a record of the calls and complaints made to the Secretary of State’s office. County elections officials are not required to maintain a record of complaints their offices receive and they are not required to report any complaints or problems to the Secretary of State’s office. The Secretary of State’s office has a formal process in place to collect, catalog, and respond to each and every complaint filed with the office on Election Day. The Secretary of State operates a toll‑free hotline that can accommodate as many as 60 calls at a time on Election Day and is staffed by Secretary of State employees and volunteers who assist voters. Each call to the hotline is recorded in a database and is tracked and responded to appropriately. On February 5, 2008, the Secretary of State’s office received over 21,000 calls from voters who needed information or wanted to file a complaint. Fewer than 2.5% of those calls led to complaints being filed. Of the 540 complaints that were filed, 96 were related to various decline‑to‑state voter issues or poll worker training issues with regard to information on how the slightly ajar primary process works in California. Thank you again for the professional fashion in which you and your staff conducted this audit, and for the opportunity to further clarify some of the issues noted in the audit report. Sincerely, (Signed by: Debra Bowen) Debra Bowen Secretary of State California State Auditor Report 2008-106 55 September 2008 (Agency response provided as text only.) Alameda County Registrar of Voters 1225 Fallon Street Oakland, California 94612 September 5, 2008 Elaine M. Howle, State Auditor Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: Attached is Alameda County’s response to the report we received based on the audit your office conducted as requested by the Joint Legislative Audit Committee. It is our goal to provide voters with a positive voting experience and to ensure the rights of all voters are protected. We plan to immediately implement the improvements outlined by your office in this report. The training guidelines are a valuable tool which we use along with training material from organizations such as the California Association of Clerks and Elections Officials, the United States Election Assistance Commission and through networking with other county election officials. These valuable resources provide a continuous flow of substantial information that we consistently use in our ongoing efforts to develop and improve training procedures for our poll worker training program. Alameda County Registrar of Voters has made noteworthy strides in the last several years. Poll worker training has been expanded and is also available on‑line. Careful attention has been given to the quality and practicality of the training. By implementing the Help Desk Electronic Automated Tracking System (HEAT) Alameda County has been able to provide superior customer service to its voters. We shall continue to diligently train and support our poll workers using the suggestions and recommendations provided by your office. If you have any questions, please contact me at 510‑272‑6933. Sincerely, (Signed by: Cynthia Cornejo for) Dave MacDonald Registrar of Voters 56 California State Auditor Report 2008-106 September 2008 Alameda County recognizes that state law requires uniform standards for training poll workers in a minimum of eight topic areas. It was identified that Alameda County does not include cultural sensitivity training in its poll worker training curriculum. We recognize the need for this component. We are developing material to address this deficiency. Our trainers provide poll workers with a booklet entitled “What to do if….” This booklet describes the role of the poll worker in assisting voters (two persons or curbside). We appreciate that the auditor acknowledged the written material, and we will ensure that it is discussed in future training. We recognize the need to continually update and revise poll worker training. We will use the Poll Worker Training Guidelines as a tool, combined with other available resources mentioned in the report. The report indicated that we need to provide role‑playing exercises and incorporate them into the training for the November 2008 election. Hands on training did occur in a one hour segment subsequent to the verbal instruction in May 2008. Every inspector had an opportunity to operate both the touch screen and the ballot scanner as a part of the training. We are currently revising our training to add the role‑playing component. We agree that in order to maximize the number of poll worker training sessions we should schedule them for different times, locations and days of the week. We feel that we don’t have to entice our poll workers to training in that a mixed class of experience is beneficial because rules and procedures change from election to election. Some of the more experienced poll workers become mentors for the first time poll workers. The ability for poll workers to refresh their skills prior to Election Day is invaluable. We are developing a video tape of the training and will have this available on‑line for inspectors and clerks to review. Additionally, county staff will set up a “mock” election officials table where poll workers can come and role play Election Day procedures/scenarios. Alameda County made revisions to the poll worker training for June 2008 after reviewing poll worker surveys and receiving election officials’ input. This is based upon the inspectors citing a need for more voting machine troubleshooting (so that they did not have to rely on the precinct coordinators as much, and could do it more quickly). We also determined that inspectors needed more instruction on opening and closing the polls on Election Day. Alameda County elections staff interview all candidates for the job of poll worker trainer prior to hiring them. County staff reviews the poll worker training lesson plans prior to the training being held. The resume of the third trainer had not been obtained because the trainer has approximately 20 years of training experience and is a retired elections official from another county.. Alameda County will continue to ensure the quality of the poll worker training by having the instructors practice teaching the material to county personnel. This has proven useful as it has allowed us to adjust content and technique prior to the poll workers receiving the training. Prior to the February 2008 election Alameda County had a manual poll worker training tracking system. Subsequently we have implemented an automated poll worker information system that records and identifies poll workers who have been to training and those who have not. We can now accurately record and track this data. All poll workers are required to attend training before working Election Day. Page 1 of 2 California State Auditor Report 2008-106 57 September 2008 Alameda County utilizes the Help Desk Electronic Automated Tracking System (HEAT). This provides us with the opportunity to respond to a variety of Election Day issues by telephone (as a help desk). Calls are logged into the program and the reports indicate how the problem was resolved. This software program has the ability to provide us with subject/category specific reports. Election officials held a post‑election critique of the February 2008 election. The results of this were used to revise training methodology for the June 2008 election. We recognize the need to document/summarize this information. Although the HEAT system identifies problematic areas and is used as the catalyst in the post‑election debriefing, we are currently working on a document that will record our summarization of voter complaints and post election debriefing. This concludes our response to your report. Page 2 of 2 58 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 59 September 2008 (Agency response provided as text only.) Fresno County Registrar of Voters 2221 Kern Street Fresno, CA 93721 Elaine M. Howle, State Auditor* Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 FRESNO COUNTY RESPONSE TO BUREAU OF STATE AUDITS September 3, 2008 The training and recruiting of poll workers for a major election is both a challenging and a rewarding process. Fresno County prides itself on providing a comprehensive training for all poll workers, inspectors and clerks. Prior to every election the ROV recruits and trains a reserve pool of thirty to forty poll workers who report to the ROV office at 5:30 a.m. on election morning to fill positions for poll workers who fail to report to work. There have been very few instances where we have filled a poll worker position with a worker who has not been trained. If an inspector position is filled, as an example, the night before the election for an inspector that is unable to work we recruit from our pool of reserves or recruit an inspector who was not assigned previously to that election but is willing to work when we call them. It is our plan to use the suggestions provided to us by the Bureau of State Audits report and will implement many of these suggestions for the November 4, 2008 Presidential Election. Fresno County has also recently produced a professional video for poll workers on awareness of accessibility issues. The theme throughout the video is “Awareness, Respect and Common Sense” in dealing with the disabled community. This video has been forwarded to the Bureau of State Audits for their review. It will be used in the training for the November 4, 2008 election and all future elections. Page 34 (2nd paragraph)—Fresno County strives to train all poll workers, (inspectors and clerks). A thorough 1 record of poll workers attending training classes is maintained in the elections management system, EIMS. 2 These records were provided to the auditors. The county did not retain the attendance cards from each training class because these cards are used to update the elections management system and the paper trail was not required. If an Inspector drops out and cannot work and notifies the county on the Monday night prior to the election, the county may need to place an Inspector who has not been trained for that particular election but who has worked in prior elections. Page 36 (3rd paragraph)—Fresno County offers hands on practice after the training class for any poll worker 1 who is unfamiliar with the use of the voting machines. The county also offers a “Lab Day” to go through the 3 set up and use of the voting machines for any poll worker who wishes to attend. * California State Auditor’s comments appear on page 61. 60 California State Auditor Report 2008-106 September 2008 1 Page 39 (2nd paragraph)—It should be noted that there has been little mention of the Secretary of State Poll Worker Training Guidelines since they first were distributed in 2006 and no direction from the Secretary of State’s office on the implementation of these guidelines. Fresno County will review these guidelines, for all future elections and will implement changes for the November 4, 2008 election. (Signed by: Kathy McClue) Kathy McClue Assistant Registrar of Voters County of Fresno California State Auditor Report 2008-106 61 September 2008 Comments CAlIfORnIA STATe AudITOR’S COMMenTS On The ReSPOnSe fROM fReSnO COunTy To provide clarity and perspective, we are commenting on the response to our audit report from Fresno County. The numbers below correspond to the numbers we have placed in the margin of the county’s response. While preparing our draft report for publication, page numbers 1 shifted and therefore the page numbers the county cites may not correspond to the page numbers in our final report. Fresno County did not provide us a thorough record of poll 2 workers attending training classes. As we state on page 35 of the audit report, it did not have training records for the February 2008 election and its training records for the June 2008 election were incomplete, with six of the 29 trained poll workers in our sample missing from the training list provided. Fresno County did not offer hands‑on training during the training 3 class we observed. We reflect this conclusion in Table 2 on page 28 of the audit report. 62 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 63 September 2008 (Agency response provided as text only.) Kings County Registrar of Voters 1400 West Lacey Blvd Hanford, CA 93230‑5905 September 5, 2008 Elaine M. Howle, State Auditor Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: As the Kings County Registrar of Voters I have reviewed the redacted draft copy of your report titled “County Poll Workers: The Secretary of State’s Office Has Developed Statewide Guidelines, but County Training Programs Need Some Improvement”. Below is my response. As you may know, Kings County is a small rural county located in the Central San Joaquin Valley. Our staff of six is responsible for the registration of nearly 45,000 voters and the administration of elections for all Federal, State, County, City, School District and Special District elections. Under the best of circumstances the administration of elections is a challenging job but events of the past two years have made this challenging job all the more difficult. In March of 2007 the Governor and the State Legislature separated the Presidential primary from the election normally held in June moving the Presidential primary forward to February and leaving State and Local elections to the primary in June. In August of 2007 the Secretary of State decertified the electronic voting system used in Kings and 26 other counties throughout the State. Prior to decertification, Kings county used a popular electronic touch screen system for all voters who chose to vote at one of our 39 polling places in Kings County. The addition of another election and the need to completely change voting methods, develop new voting procedures and recruit and train election officers for a non‑traditional election day added to the challenges of this past year. Through all this, Kings County has been lucky to have long serving and dedicated election day officers who put in long hours and who are dedicated to making the voting process as efficient and enjoyable as possible for all voters. The majority of our polling place workers have been with us for many years. We recently recognized one worker who has served with us for the past thirty years. While this is an admirable record, decades of service is not unusual among our workers. In light of the ever changing requirements on both our office staff and election day poll workers I welcome any help and direction that can be derived from the process that produced this draft report. and while I respectfully disagree with some of the conclusions of this report the process has been enlightening. I look forward to reading the final report in its entirety. Sincerely, (Signed by: Ken Baird) KEN BAIRD Kings County Assessor, Clerk/Recorder, Registrar of Voters 64 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 65 September 2008 (Agency response provided as text only.) County of Los Angeles Registrar of Voters 12400 Imperial Highway – P.O. Box 1024 Norwalk, CA 90651‑1024 September 4, 2008 Elaine M. Howle, State Auditor* Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: Thank you for conducting a review of pollworker training and how counties address Election Day complaints regarding the voting experience as identified in the audit goals and the observations covered in this report. We found that your staff was sensitive, interactive and respectful. They sought to gain a true insight into the subject matter covered in the audit scope under challenging circumstances – reviewing, in a short time frame, some of the obscure and intricate mechanics behind conducting elections in the largest state in our nation. This work helps provide much needed information regarding the complexities that counties face with regard to: • the recruitment, retention and training of pollworkers; • managing and tracking unusual incidents at poll places on Election Day; and • navigating the rules and variables inherent in California’s unusual Primary election system. It also points to the enormous volume of election activities in Los Angeles County related to its uniquely large number of registered voters and polling places – the largest in the United States. We appreciate the emergent perspective that no two counties can or do conduct an election exactly alike given the diversity of the state and the specific needs related to each county’s unique circumstances and voting systems. We were pleased that the report specifically pointed out that Los Angeles County complies with the content of the Secretary of State’s training guidelines when training Inspectors, as well as other pollworkers, and recognition that there are noteworthy training practices taking place in Los Angeles County. There are some areas of the audit where we will feel obliged to comment in order to offer further context and perspective with regard to Los Angeles County’s practices and programs. That commentary follows with reference to particular headings or sections in your audit report. * California State Auditor’s comments begin on page 71. 66 California State Auditor Report 2008-106 September 2008 Actions counties take when they receive complaints from voters This topic, with regard to Los Angeles County, is covered toward the end of the report’s Introduction as well 1 as in Chapter 2 under the section titled “Only One County We Reviewed Kept Data On How It Responded to Complaints” (Los Angeles County is that county). We appreciate the implicit recognition of the enormous Election Day incident/troubleshooting effort that Los Angeles County managed for the February 2008 Presidential Primary Election and the related tracking system and database described in these two sections. An effort of this kind and magnitude—designed to protect voters’ rights by providing for quick, documented responses to polling place incidents—may be unique and is an effort that Los Angeles County has undertaken for years. Your review of the system and implicit recognition that we collect and maintain voluminous data regarding Election Day complaints within the mechanism of the system is significant to our overall operations. We are concerned, however, with the implication in your Introduction that the system may have serious data integrity issues. We are similarly concerned that the characterization in the two sections could indicate, given 2 an idiosyncratic reading, that the data in the former system may have been intentionally mishandled. The data selection and analysis noted in the report is concerning given that after we were provided with figures that 27 out of 29 complaint forms in your sample did not agree with source documents, we performed an analysis that revealed that at least 423 of the intake forms exactly matched that which was in the electronic data. Our assessment is that a more reasonable summary of the data related to the system would be that there was a statistically significant discrepancy of the sample input forms when compared to the electronic data. As a result, Los Angeles County’s’ complaint tracking data was not sufficiently reliable for 3 the purpose of evaluating how the County responds to voter complaints. We also feel that there is a more reasonable characterization of information in Chapter 2 under the section 1 titled “Only One County We Reviewed Kept Data On How It Responded to Complaints” (Los Angeles County) given the concern previously stated. That is, we feel that the sentence in that section that reads: “Los Angeles County’s ability to use the information in its complaint database to determine how effectively and promptly it resolved complaints is limited primarily because its data is inaccurate and incomplete” is more 4 accurately characterized by stating the audit team was unable to draw conclusions about the effectiveness of this data in post election analysis since the data is inconsistent for auditing purposes. For example, Los Angeles County did not enter all complaint forms into its database; therefore the database was not a comprehensive resource for assessing responses to complaints. Additionally, we feel that it is important to identify and make reference to the practical objectives of the system used during the elections reviewed: • To quickly document poll place incidents and resolve them on Election Day without letting overly forma documentation practices slow the rate of solving real election problems in the field. (That is, staff acted under the general rule that they should document as precisely as possible but not to the detriment of solving the next issue in cue); and • To provide a method whereby experienced staff could – in a calm period following Election Day—examine intake forms and conform them to a known typology of issues that would be meaningful for specific follow up and general trend analysis. 2 California State Auditor Report 2008-106 67 September 2008 In summary, the system was designed primarily to address voter issues in real time on Election Day and to provide a reasonably reviewable body of data after the election. Absolutely precise data collection was not the purpose of the system. It is important to indicate that we speak of the system in the past tense ‑ since it was replaced prior to the 5 June 2008 Primary Election by a state‑of‑the‑art, automated Help Desk system that addresses all of the items implicitly critiqued by your review as noted later in your report. Chapter Two Summary/ Not All Poll Workers Are Required to Attend Training And Most Counties We Visited Could Not Provide Reliable Training Data In these two sections, it is either stated or implied that Los Angeles County was one of two counties that provided reliable training attendance data but also did not require all Inspectors to attend training prior to working during the February 2008 election. Los Angeles County’s policy is to require all Inspectors to attend training class. However, as you indicate in the report, a number of Inspectors served when they did not attend in‑person training classes for the given election. Although the causes of this phenomenon are often due to extenuating circumstances such as Inspectors resigning on short notice before Election Day, hard‑to‑recruit‑areas, the uniquely large number of Inspectors recruited by Los Angeles County, etc., it is a number that we are constantly seeking to decrease. That said, instructional materials and other resources are, in fact, provided to ALL Inspectors. The issue of Inspectors who do not attend an in‑person training class has been one that the Department has been addressing for some time by seeking to mitigate its impact on Election Day and by using methods to drive the number down as quickly as possible. For example: • To mitigate effects on Election Day, we have intentionally created written pollworker materials—as well as video materials—to ensure that pollworkers have reference information adequate to run a poll “from scratch” on Election Day. Indeed, it is a core philosophy of our training program that pollworkers should rely on their written materials as a check list and/or troubleshooting guide which will greatly help them in navigating the complexities of Election Day. • We have also greatly increased the number of coordinators to assist pollworkers in the field in recent years—providing onsite orientation and training to late replacement Inspectors. • In recent years, we provided a significant monetary incentive that resulted in a large number of clerks attending training class compared to attendance before the incentive was introduced. During the course of this training, clerks are trained on much of the same material that Inspectors receive in their training classes with the intent that ‑ collectively ‑ they can pool their knowledge especially when an Inspector may be experiencing specific challenges. • We have also seen a significant decrease in the number of Inspectors who do not attend in‑person training related to the recent increase in the monetary incentive to attend class. 3 68 California State Auditor Report 2008-106 September 2008 We will continue our efforts—in earnest—to reduce the number of Inspectors who do not attend in‑person training prior to each election. Please note that the number of pollworkers that did not attend in‑person training indicated in the 6 report (317) has gone down and will continue to go down as we receive communications from pollworkers who attended training but who were not credited for doing so. 1 Chapter Two: Table 3‑ Methods Used to Collect Data on the Effectiveness of Poll Worker Training After much interaction with the audit team regarding your finding of “No” in the column, Evaluations of 7 Training by the Instructors Teaching the Class, we still have a significant concern about this characterization and we believe the finding does not reflect our work in this regard. Our concern is related to the possible perception that we do not allow our staff to evaluate and assist in the development of our pollworker training programs, when, in fact, we rely on and encourage them to provide a significant amount of feedback and input into those programs. The County employs 5 full time staff in a work unit that is almost exclusively dedicated to developing pollworker training year round. In addition, over the past year, at least two – and at times three—experienced seasonal employees have been part of this work unit for at least six months at a time who have worked with other seasonal training staff during previous elections. These fulltime and semi‑fulltime ”training development” workers are involved in day‑to‑day activity that is solely focused on poll worker training—activity devoted to subjecting the next round of training to a continuous improvement process that manifests itself in the class curriculum. This activity includes regular one‑on‑one and group debriefing/feedback sessions before, during and after the training program for any given election.: Additionally, other County staff including the Assistant Registrar‑Recorder/County Clerk in our Election Services Bureau consistently receives and gives feedback to our trainers. During the audit, we provided 7 a great deal of information that specifically described and attested to such feedback opportunities. 8 Additionally, the audit team experienced first hand a small example of one of these opportunities when they observed a post‑training debriefing of trainers. During that session, the trainers were asked to report out on their own perception of their training delivery and the program in general. It is important to indicate that, like most continuous process improvement activities that involve professional full time staff, copious documentation is not kept on all meetings, debriefings, and conversations regarding producing a better product; however the results of this interaction are seen in the final training products that change from election to election despite the seeming implication that such documentation is obligatory. We understand that your agency has technical reasoning related to methodological considerations 7 regarding the “No” finding but they remain opaque to us and leave us unclear as to what practices, short of formally documenting each staff interaction, would have merited a “Yes”. Regarding, the footnote to your “Yes” characterization in the column, Electronic‑based Evaluation of Online Training by Poll Workers, in our initial “roll‑out” of on‑line training, in order to not overwhelm our pollworkers and other resources, we limited our formal survey. That is, pollworkers who are County employees were given the option of taking the online training and submitting a survey. A small number of County pollworkers completed that survey. 4 California State Auditor Report 2008-106 69 September 2008 County staff also did informal interviews of a number of pollworkers regarding the online training experience. We believe that by requiring coordinators to pass an online assessment test in order to serve on Election Day, that we garnered an implicit survey of the system that provided us definite data patterns. Our long‑term plan remains to survey a large body of our pollworkers and coordinators regarding the online training experience on an ongoing basis. None of the Eight Counties Could Demonstrate How They Identified Needed Changes to Poll Worker Training We believe the title and some of the content of this section are potentially misleading given that we 9 submitted significant evidence to your team of changes to our training program between the February and June 2008 Primary Elections including a description of those changes that was submitted to our Board of Supervisors and Chief Executive Officer on July 10, 2008. Many of those changes were created with the assistance of our well‑respected community partners/stakeholders who are described on page 4 of the July 10 report. Most concerning about the title of this section is that the report indeed describes the fact that 9 Los Angeles County identified and made needed changes to pollworker training notwithstanding the opposite indication stated in the title. We believe it is relevant to note that the sample studied – two Primary Elections held within months of each other (something that had not occurred in over fifty years) – was not adequate to identify the kind of continuous and significant improvement that you would have identified had you added additional elections into the mix for comparison. Again, thank you for the effort your office put forth in helping create a better understanding of the election process in California and for this opportunity to respond to your report on these matters that are extremely serious in the context of our democracy. Sincerely, (Signed by: Dean C. Logan) DEAN C. LOGAN Registrar‑Recorder/County Clerk 5 70 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 71 September 2008 Comments CAlIfORnIA STATe AudITOR’S COMMenTS On The ReSPOnSe fROM lOS AngeleS COunTy To provide clarity and perspective, we are commenting on the response to our audit report from Los Angeles County. The numbers below correspond to the numbers we have placed in the margin of the county’s response. While preparing our draft report for publication some wording 1 changed, including the section title the county refers to. The audit report does not state that county staff intentionally 2 mishandled the complaint data. On page 13 of the report, we discuss our conclusion that the data in Los Angeles’ complaint tracking system was not sufficiently reliable for the purpose of evaluating how the county responds to voter complaints. We appreciate that Los Angeles County acknowledges in its 3 response that its data was not sufficiently reliable for the purpose of evaluating how it responds to voter complaints. We believe our characterization of the condition of Los Angeles 4 County’s complaint database on page 43 of the report is accurate. Therefore, we have not altered the text in the final report. The system we reviewed during fieldwork was the system the 5 county used during the February 2008 election, which was the most recent election at the time we conducted our fieldwork in May 2008. We recognize that the system we reviewed during our audit fieldwork has since been replaced and we provide the county’s perspective on its new system on page 44 of the final report. We cannot comment on the accuracy of the county’s claim that 6 the number of untrained inspectors has decreased from 317. As we report on page 36 of the report, in July 2008 the county asserted that 3,965 out of 4,282 inspectors attended in‑person training for the June 2008 election. The number 317 represents the difference between these two numbers. We stand behind our characterization of the results of our audit 7 work in Table 3 of the report. As we note in Table 3 on page 37, we required counties to provide documentary evidence of their data collection practices in order to support a positive rating. Audit standards require that we obtain sufficient and appropriate evidence to support our conclusions. Audit evidence can take the form of an auditor’s direct observation, documentary evidence, or testimonial evidence. In our judgment, relying on the county’s 72 California State Auditor Report 2008-106 September 2008 testimonial evidence in this instance, without the benefit of corroborating documentary evidence, would not comply with audit standards and would not provide a sufficient basis for our conclusions. Finally, our report does not conclude that counties fail to employ the methods shown in Table 3. Instead, our point is that we did not always see evidence of these methods in practice. 8 Los Angeles County’s response suggests that our auditors had direct observation of instructors providing feedback on the training class we observed, and that we therefore should have provided a positive rating in Table 3 on page 37 of our report under column three relating to direct feedback by instructors. However, the discussion the county refers to was not initiated and facilitated by the instructors teaching the class, but rather by the county’s assistant registrar who accompanied us to the training. Therefore, we stand behind our characterization of this issue in Table 3. 9 We disagree with the county’s claim that the title of this section is inconsistent with the text that supports it. As we state on page 40, Los Angeles County was unable to provide us with any documentation of its analysis of the February 2008 election until late August 2008, nearly two months after the June 2008 election. While it is true that we saw significant changes to its June 2008 training program with respect to decline‑to‑state voters, such an observation does not alter the fact that Los Angeles County could not demonstrate to us how the data it collected was used to identify needed changes to its poll worker training. In fact, as we state on page 39, at the time of our fieldwork Los Angeles County indicated that its process for updating poll worker training was not always documented. California State Auditor Report 2008-106 73 September 2008 (Agency response provided as text only.) Orange County Registrar of Voters 1300 South Grand Ave, Bldg. C Santa Ana, CA 92705 September 5, 2008 Elaine M. Howle, State Auditor* Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 RE: Orange County’s Response to California Bureau of State Audits Report on County Training Programs Dear Ms. Howle: Enclosed you will find the response from Orange County to the Bureau of State Audits (BSA) review of county poll worker training programs. I would like to commend the staff at the BSA for their professionalism and willingness to work with us to provide the best assessment possible of poll worker training. Please do not hesitate to contact me should you need further information for your report. Sincerely, (Signed by: Neal Kelley) Neal Kelley Registrar of Voters * California State Auditor’s comments appear on page 77. 74 California State Auditor Report 2008-106 September 2008 Orange County Response to California Bureau of State Audits Report: “County Poll Workers: The Secretary of State’s Office Has Developed Statewide Guidelines, but County Training Programs Need Some Improvement” The Bureau of State Audits (BSA) stated that many of the eight counties they reviewed look to other sources of 1 information to update their training programs. (page 5) These sources include the CACEO, the federal Election Assistance Commission or other counties. Although Orange County is not listed as one that principally utilizes these sources to change its poll worker training it should be noted that many of the topics and techniques used by these sources are taught and used in Orange County. Although Generally Consistent With Guidelines For Training Poll Workers, Most Counties’ Training Programs Need Improvement The BSA states that not all counties, including Orange County, were able to provide data that demonstrated 1 they trained all inspectors prior to the February 2008 election. (page 34 & 47) Orange County understands that the BSA decided to not accept evidence of inspectors that attended training prior to the February 2008 election due to a notation from the Registrar employee 2 responsible for the data who believed that we needed to take a closer look at the data. The county respects the opinion of the BSA, however, it is the opinion of the county that the note and the memo as a whole provided no proof that this was actually the case and believe that further review of the data would show that the inspectors that worked in Orange County in February 2008 did, in fact, attend training. None of the Eight Counties Could Demonstrate How They Identified Needed Changes to Poll Worker Training The BSA states that none of the counties could show a document that linked conclusions from poll worker survey data to proposed changes in training. 1 (page 52) The post election reports, including debriefing documents provided to the BSA include information that was gathered from those surveys and phone calls taken on Election Day. Both sources indicated that poll workers would benefit from having different levels of training. The County of Orange changed its training from February 2008 to June 2008 by providing separate classes for new clerks, experienced clerks and Inspectors. This information can be linked by reviewing the poll worker surveys, 3 feedback from poll worker trainers and debriefing documents. A debrief from a poll worker trainer that also worked on the poll worker technical support line on Election Day, as all trainers do in Orange County, made a specific recommendation regarding the need for enhanced training for canceling an access code versus closing the polls on the electronic voting system. Information regarding poll worker confusion was documented on the database used by the phone bank. The training manual for the June 2008 election has changes in language for these two areas to better describe the difference. The PowerPoint was also changed between these two elections to reflect the enhanced training in this area as was the hands‑on portion of the class.: The BSA says that it did not appear that Orange County took voter feedback into consideration where poll workers “were not adequately informed or misunderstood election procedures”. 1 (page 54) Although no specific list was provided with feedback from voters, we would like to note that we have a) Public phone bank taking comments, questions and complaints b) Poll Worker Phone Bank where 1 4 California State Auditor Report 2008-106 75 September 2008 calls are recorded that include issues that the poll workers are having, including questions that come from voters c) Rapid Deployment Teams that are loaded with equipment and technical expertise that are dispatched to polling places in need of assistance d) A Community Outreach Manager (also manages training) that answers media, community group, political party, Secretary of State, poll worker and voter phone calls on Election Day while monitoring the poll worker customer service line database e)A Registrar of Voters that trouble shoots issues at poll sites during the majority of Election Day. Members of each of these areas are involved in a debriefing meeting where various topics are discussed and changes to be made are addressed. Although documents provided to the BSA might not refer directly to voter calls, it does not mean that they are not reflected. Only One County We Reviewed Kept Data On How it Responded to Complaints 1 The BSA states that counties discussed procedures for handling voter complaints and poll worker complaints, but that none were able to provide data that was both reliable and described how a complaint was resolved. (page 56) The County of Orange has a Poll Worker Technical Support Line set up on Election Day that 1 includes a database for tracking poll worker questions, comments and complaints. The BSA was provided with a summary of the main issues addressed in February that was provided to the Secretary of State. This was a summary document and we understand that the BSA would like to have seen a more comprehensive document. In the future we will develop a more detailed breakdown of potential issues. The database includes a section for how the call was resolved. After February 2008 the County of Orange established a Poll Worker Customer Service Line that is open 60 days prior to the election. The customer service representatives are available 11 hours a day for five days a week to answer any poll worker questions. The calls are recorded in a database and the customer service representatives work closely with the poll worker recruiters. Conclusion The Orange County Registrar of Voters office would like to thank the BSA for their professionalism in dealing with us during this process. We recognize that a majority of the recommendations found in their report are part of Orange County’s training program and we are dedicated to continual improvement. In fact, all recommendations will be implemented in Orange County for the November 2008 election, including a dedicated web page for poll workers.: 2 76 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 77 September 2008 Comments CAlIfORnIA STATe AudITOR’S COMMenTS On The ReSPOnSe fROM ORAnge COunTy To provide clarity and perspective, we are commenting on the response to our audit report from Orange County. The numbers below correspond to the numbers we have placed in the margin of the county’s response. While preparing our draft report for publication, some wording 1 changed. In addition, page numbers shifted and therefore the page numbers the county cites may not correspond to the page numbers in our final report. We appreciate that Orange County respects our decision to use its 2 own internal analysis of the February 2008 election to conclude that its training data was inaccurate. Orange County’s “Lessons Learned from February Election” document states that “attendance [records] is not 100 percent accurate” and “if only a group of people would [have] taken care of the rescheduling it would have been so much easier to jot down the non‑attendees who missed training and of course to track down all documentation on these individuals.” During our exit conference, county officials expressed frustration with our conclusion and submitted an altered version of its “Lessons Learned from February Election” document with some of the above quoted statements removed. Thus, we are skeptical of the county’s opinion that its data are accurate. We stand by our conclusion that Orange County did not link its 3 conclusions about how poll worker training should be improved to any data or analysis it performed. As we note on page 40 of the report, Orange County’s “Lessons Learned from February Election” document for the February 2008 election listed several recommended changes based on surveys of instructors and county staff. However, the tabulated results from these surveys were not mentioned in the document. As such, the linkage to the surveys and other information Orange County speaks of is not clear. 78 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 79 September 2008 (Agency response provided as text only.) County of San Diego Registrar of Voters 5201 Ruffin Road, Suite I San Diego, CA 92123‑1693 September 5, 2008 Elaine M. Howle, State Auditor* Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: This is in response to your August 29, 2008 letter and the enclosed draft copies of the audit report which the Joint Legislative Audit Committee asked you to prepare. It is our pleasure to respond to the draft audit report. We would also like to take this opportunity to thank Wesley Opp and Mark Needham of your staff for their courtesy and professionalism during the course of this review. The focus of this response is two‑fold: First, we offer comments regarding the basis of the audit itself, the timing of the undertaking, and its general tone. Second, we address comments made in the report about the specifics of poll worker training in San Diego County in order to clarify the characterizations presented. General Comments about the Audit, its Nature, Timing, and Tone No Statutory or Regulatory Authority Exists We believe the very concept of an audit of county poll worker training programs relies on a weak premise of 1 statutory authority. It is true Elections Code section 12309.5 requires the Secretary of State to establish a task force to study and recommend uniform poll worker training guidelines, and indeed such a task force was convened. However, the recommendations of the task force do not appear to have ever been adopted by the Secretary of State. In fact the Word version posted on the Secretary of State’s website still contains track changes and the PDF version mentions “draft1” in the document title. Nevertheless the guidelines offer reasonable suggestions—due in large part to the role of county elections officials in drafting them—and San Diego County takes no issue with them. But, the fact remains, that they 2 are simply guidelines and nothing more. It is therefore questionable that any governmental or oversight body has the authority to “audit” counties’ conformance or lack thereof to a concept or guideline that has not been formally adopted as a uniform standard by the Secretary of State. Legislature’s Role in February Problems Remains Unexamined The audit strikes us as misplaced for yet another reason. The Legislature called for the audit in the aftermath of the February 2008 Presidential Primary election when many voters were confused about their right * California State Auditor’s comments begin on page 83. 1 80 California State Auditor Report 2008-106 September 2008 to vote a particular political party’s ballot. However, the reason for the confusion rests primarily with the 3 Legislature itself. The Legislature enacted the “modified open primary” in 2000, after the U.S. Supreme Court invalidated the state’s blanket primary which voters had wholeheartedly embraced. By contrast with the blanket primary, the modified primary system was a less desirable alternative and was not especially well understood by the electorate. Only “Decline to State” voters benefited, and then, only when a particular party decided to allow them to participate. Since 2000, the Republican Party has allowed Decline to State voters to participate in their Direct Primary elections but not in their Presidential Primary elections. This was a little‑noticed fact for most Decline to State voters because, until 2008, the Direct Primary and the Presidential Primary have always been combined. The net effect for Decline to State voters was that they could receive a Republican Party ballot. The fact that it did not contain a contest for President went largely unnoticed in 2004 when an incumbent President was seeking the nomination. It is against this confusing backdrop that voters in the February Presidential Primary went to the polls. Despite advance warning and ample poll worker training, the effect of the bifurcated primary election was to deprive Decline to State voters of the Republican Party ballot they were accustomed to voting and to generate widespread confusion about the nature of the modified primary system in general. The process was simply not intuitive for voters or poll workers. 3 We believe it is unfair that the report omits any reference to the Legislature’s role in the situation and attributes the confusion entirely to a failure on the part of county election officials and their volunteer poll workers. Accusatory Tone Belies Content of Report 4 We observe that the tone of the report is established by a headline which negatively describes county compliance with the guidelines: “COUNTY POLL WORKERS: The Secretary of State’s office has developed statewide guidelines, but county training programs need some improvement.” We believe the title should reflect what the report actually says, which is that counties generally comply 4 with the guidelines. This characterization is reflected in every table presented in the report where, by far, the majority of columns are marked “yes.” We urge you to restate the headline in more positive terms warranted by the body of the report. A more accurate title would read: 4 “COUNTY POLL WORKERS: State and local officials’ joint efforts to craft poll worker training guidelines have garnered support and led to considerable innovation.” Timing of Audit Detrimental to County Participation Finally, we object to the timing of the audit and specifically the fact that it was launched just days before the June election. Because election officials were at the peak of their workload, the opportunity to respond to the audit stretched resources beyond reasonable measure. Now, the release of the draft report has occurred at the upswing of yet another peak workload period just 60 days before the largest election of the four year cycle. We believe the fact that we have been given only five days to comment severely limits the ability of 5 county elections officials to respond effectively. 2 California State Auditor Report 2008-106 81 September 2008 Response to Specific Comments Regarding San Diego County Training Page 5: Reference to other agencies’ poll worker training documents 6 Although we do consult other training documents, it remains the case that San Diego County is uniquely well‑positioned to write its own poll worker training materials based on the hands‑on experience of staff. This experience is lacking on the part of other organizations. While well‑meaning, the Election Assistance Commission and Secretary of State do not directly administer elections and do not train poll workers. As such, they are inexperienced in the finer points of procedures relevant to our county, our foreign language requirements, and our voting system as deployed in San Diego County. Pages 6 and 48: Reference that not all counties were able to provide data that demonstrated that they 6 trained all inspectors prior to the February 2008 election and as a result, the counties cannot be certain to what extent these workers who supervise polling places have the knowledge to administer elections In San Diego County all three Inspector positions must attend training. All clerks are also invited to class and 50% of them elected to attend in February 2008. The class sign‑in rosters are the County’s actual 7 documentation that a poll worker attended training. In addition, Precinct Inspectors and Touchscreen Inspectors cannot pick up their election supplies without a proof of training stamped, signed and dated on their supply receipt. This official supply receipt is further evidence that these inspectors are trained prior to the election. San Diego County’s mission statement for training is “to ensure poll workers understand and complete their Election Day responsibilities with accuracy, confidence and graciousness by providing interactive training and supportive learning resources.” As outlined in this report, San Diego’s commitment to preparing poll workers with the knowledge to administer elections is clear through the hands on exercises in class and the additional resources offered (e.g., five days of workshops which poll workers may attend at their convenience, 10 days of a pre‑election hotline for poll workers to call with any questions, and the online training program to review and test their knowledge on election procedures). Pages 7 and 54: Reference that election evaluation reports did not link their conclusions from the data 6 collected to the proposed changes to be made There is an intensive debriefing process that follows each election. Through a series of staff meetings, the nature of voter complaints as well as voter, poll worker and staff recommendations for improvements are discussed thoroughly and culminate in a summary debriefing report. Because the reasons for improvements are discussed in detail in staff meetings that last between 3 – 16 hours, the document that summarizes the changes between each election does not correlate an exact complaint or recommendation to the change that was made. Thank you again for this opportunity to respond. If you have any question or concerns, please do not hesitate to contact me directly at (858) 694‑3401 or my Chief Deputy for Election Services, Nicole Alejandre, at (858) 495‑5492. Sincerely, (Signed by: Nicole Alejandre for) DEBORAH SEILER Registrar of Voters 3 82 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 83 September 2008 Comments CAlIfORnIA STATe AudITOR’S COMMenTS On The ReSPOnSe fROM SAn dIegO COunTy To provide clarity and perspective, we are commenting on the response to our audit report from San Diego County. The numbers below correspond to the numbers we have placed in the margin of the county’s response. San Diego County’s comments questioning the statutory authority 1 for the audit are unclear. The county seems to question both our authority to do the audit and the applicability of the Office of the Secretary of State’s (office) poll worker training guidelines. Government Code, Section 8546.1(b) states, in part, “[t]he State Auditor shall conduct any audit of a state or local governmental agency or any other publicly created entity that is requested by the Joint Legislative Audit Committee.” On March 12, 2008, the Joint Legislative Audit Committee approved the audit request and directed the State Auditor’s Office to conduct the audit. Further, the office informed all county registrars on April 19, 2006, that its poll worker training guidelines were final and were posted on its Web site. We saw no evidence to support the county’s claim that the office’s guidelines are still in draft form. Election Code, Section 12309(a) states, “Following the appointment 2 of members of precinct boards, the elections official shall instruct inspectors so appointed concerning their duties in connection with the conduct of the election, which instruction shall conform [emphasis added] to the uniform standards adopted by the Secretary of State pursuant to Section 12309.5.” According to our legal counsel, the use of the words “shall conform” indicates a legislative intent that the training counties provide to inspectors conform to the uniform standards. Further, our legal counsel concludes that it appears that the Legislature did not intend for the Secretary of State to go through the typical rulemaking process required when state agencies adopt regulations. Instead, the legislation calls for the establishment of a task force of experts who were required to make their recommendation available for public review and comment prior to their submission to the Secretary of State and the Legislature. Thus, while San Diego suggests that the uniform standards were not publicly vetted, the task force was required by law to do so. Nonetheless, if a court of law found that the uniform standards should have gone through the typical rulemaking process, the Secretary of State could seek to have the uniform standards approved under that process. 84 California State Auditor Report 2008-106 September 2008 3 San Diego County’s response asserts that the Legislature is to blame for voter and poll worker confusion regarding the voting options available to decline‑to‑state voters, stating that it is unfair to attribute blame to county elections officials and volunteer poll workers. We disagree with San Diego County’s perspective. As we state on page 9 of the report, the Elections Code makes county elections officials responsible for training poll workers. The logical implication is that they also need to have a clear understanding of voters’ rights so they can properly train poll workers and effectively assist voters. 4 San Diego County is incorrect in suggesting that the title of the report is inconsistent with the report’s conclusions. Our report acknowledges that the eight counties we reviewed generally complied with the office’s training guidelines. The first section heading on page 26 in Chapter 2 of the report makes this point. However, as illustrated in Tables 1 and 2 on pages 27 and 28, as well as Table A in the report’s Appendix on page 52, there are areas in which counties can improve their compliance with the training guidelines and with the office’s training methods and timing suggestions. 5 San Diego County indicates that having only five days to comment on the draft report severely limited its ability to respond effectively. Our office has conducted hundreds of audits where we provided entities a five‑day period to respond and this has been our standard practice for years. Further, we read the sections of our draft report that related to San Diego County to the county’s elections officials in late August and discussed our findings and conclusions about the county with them at that time. 6 While preparing our draft report for publication, page numbers shifted and therefore the page numbers the county cites may not correspond to the page numbers shown in our final report. 7 Although San Diego County claims that poll workers who occupy any of its inspector positions must attend training, we cannot comment on the accuracy of the county’s assertion. On page 35 of the report, we state that San Diego County provided us with an incomplete training list. As a result, we concluded that San Diego County’s training records were unreliable for the purpose of evaluating whether all inspectors received training. California State Auditor Report 2008-106 85 September 2008 (Agency response provided as text only.) County of Santa Clara Registrar of Voters 1555 Berger Drive, Building 2 San Jose, CA 995112 September 5, 2008 Elaine M. Howle, State Auditor* Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: Thank you for the opportunity to review and comment on the draft report “County Poll Workers: The Secretary of State’s Office Has Developed Statewide Guidelines, but County Training Programs Need Some Improvement”. Attached are our comments and corresponding attachments.† Sincerely, (Signed by: Elaine Larson) Elaine Larson Assistant Registrar of Voters * California State Auditor’s comments begin on page 91. † We have not included attachments in the report; however, they are available for reveiw at the California State Auditor’s office. 86 California State Auditor Report 2008-106 September 2008 County of Santa Clara Response to Draft “County Poll Workers: The Secretary of State’s Office Has Developed Statewide Guidelines, but County Training Programs Need Some Improvement” September 5, 2008 1 Audit Issue on Page 34, Paragraph 2: “However, not all counties… (redacted) were not able to provide data that demonstrated they trained all inspectors prior to the February 2008 election. As a result, the counties cannot be certain to what extent these workers who supervise polling places have the knowledge to efficiently administer elections.” Santa Clara County Response: Santa Clara County disagrees. The auditors first asked for a list of all people that attended training, whether or not they actually worked. We provided the report. Later, the auditors asked for a list of people who worked and attended training. Although this may seem like the same information, it is not. Our data management system does not track people who attend training and do 2 not work. This resulted in two distinct sets of data resulting in discrepancies. We offered the state auditor an explanation for the discrepancy, but it was rejected as “unreliable”. The auditors insisted that they only wanted data they could manipulate rather than understanding our process of verifying attendance. Unfortunately, our information management database does not generate complete reports. It is necessary therefore to export data in order to provide the information. Regarding Precinct Inspectors, to our knowledge there was a difference of one. We confirmed that one inspector had in fact attended a training class, was paid for attending a training class, but we neglected to 2 go back into our data management system and enter her name into the class she attended, resulting in a discrepancy between the two different reports. This was brought to the attention of the auditors, but was rejected as unacceptable and therefore deemed “unreliable”. 3 Santa Clara County takes pride in ensuring that all Inspectors attend training or they do not work. As we 4 have tried to explain previously, all training classes have a pre‑printed class roster with names of scheduled participants on which trainees must sign. The completed rosters are submitted to the Precinct Operations analyst, who reconciles the lists with the database. If a person did not attend the class, their name is removed from the class roster and an email is sent to the respective Election Specialist to call and reschedule the person. The analyst monitors each poll worker and provides reports to the Precinct Operations managers daily. If an Inspector has not attended training by the Saturday before the election, the person is replaced by another person who has attended training or by one of the Election Officer Training Assistants. 5 We explained this information on a conference call with the State Auditor’s Office on September 4, 2008 and emailed additional summary reports that afternoon. 1 Audit Issue on Page 35, Paragraph 1: “And while all eight counties told us they receive complaint calls from voters or poll workers on Election Day, only three were able to provide us with a complaint log detailing those calls for either the February 2008 or June 2008 elections.” Santa Clara County Response: We disagree with the audit findings. Santa Clara County provided the 6 auditors with all Voter Feedback forms completed for the February and June 2008 elections. That was our log. The manager for the Voter Registration Division also provides a weekly summary of these forms to the 1 California State Auditor Report 2008-106 87 September 2008 Registrar. Copies of the telephone summary call reports were provided to auditors in May and emailed on September 4, 2008. Submitted in May to the audit teams were copies of the reports on ROV website email complaints, Election Officer and Voter Summary reports. After a Voter Feedback form is completed by a staff member, it is given to the Voter Registration manager. The Voter Registration manager reviews the forms for resolution, follow‑up and dispersion. Forms that need to be followed‑up on by other managers are noted with the manager’s name, copied and delivered. Managers who receive these forms are to record their follow‑up/resolution and return the copy to the Voter Registration manager. Forms that are just feedback or are resolved at the first contact point are copied and given to the appropriate division manager for their records. All forms received each week are assigned a “division” or category and are reported to the Administrative Support Officer who updates the summary report and presents it at the weekly operations meeting with the Registrar of Voters. Audit Issue on Page 43, Paragraphs 2 and 4: “Unfortunately, six of the eight counties we visited were unable 1 to provide reliable data to demonstrate that all of the inspectors received training prior to the February 2008 election. (redacted). However, with many of the counties having poor data on training attendance, it is uncertain to what extent poll workers have the knowledge to effectively administer elections.” “Many counties we visited were unable to provide reliable poll worker training data, which prevented us from evaluating whether all inspectors were trained prior to the February 2008 election.” Santa Clara County Response: See above response to issue Audit Issue on Pg 34, paragraph 2. We believe 2 the information was accurate and reliable. Audit Issue 3 on Page 48, paragraph 2: “(redacted) Santa Clara, (redacted) counties also had difficulty 1 providing us accurate or consistent training lists for their poll workers. (redacted). Santa Clara County provided two separate lists for February 2008 election, both prepared after the election, however, we noted inconsistencies between the two reports and the county was unable to explain these differences.” Santa Clara County Response: See above response to issue Audit Issue on Pg 34, paragraph 2. Additionally, 2 the report states that both of these reports were provided after the February election. The auditors first came to our county in May 2008 and at that time we provided the report as requested. Since it was requested in May, we could only provide a post‑election report for the February election. Audit Issue following Page 50, Table 3: “Methods Used to Collect Data on the Effectiveness of Poll Worker 1 Training”, No responses. Santa Clara County Response: Santa Clara County disagrees. The findings of “No” in the “Evaluation of Training by the instructors Teaching the Class” and “Evaluation of Training by County Staff” columns are contrary to the documentation provided to the auditors between the period of 5/19‑5/21 and by email on 8/22. These documents detailed the notes taken by assistant instructors, instructors and County staff 7 who participated in all of the training sessions and conducted a debriefing on 2/25/08. The audit team was provided with a copy of the notes taken at that debriefing session and offered additional information, timesheets and primary documents including handwritten debriefing notes from staff and management who attended the debriefing session. 2 88 California State Auditor Report 2008-106 September 2008 Continuing the objection, the findings of “No” in the “Summarized Data on Voter Complaints from Election Day” and “Summarized Data on Poll Worker Questions / Concerns from Election Day” columns are additionally contrary to the documentation provided to the auditors between the period of 5/19‑5/21 and 8 by email on 8/22. These documents summarized complaints by category that came in by email and over the phone. Furthermore, the audit team was provided primary documents of Election Officer Feedback Forms, Field Inspector Summary Reports and Election Day Problem Reports from the Election Day phone hotlines. These primary documents were compiled into summary reports provided to the audit team both in our office and by email. 5 See attached “EO Feedback and 299‑POLL Summary report file. 1 Audit Issue on Page 55, Paragraph 2: “…we could not identify any material change in Santa Clara’s training programs.” Santa Clara County Response: The audit team was provided several separate summary documents as well as primary and PDF materials outlining separately, the changes to the Election Officer Manual, Training 9 PowerPoint, Training Structure/hands‑on session and Online Training. Substantial changes were clearly demonstrated by adding a completely new module to online training, adding new hands‑on role‑play sections to class. The Roster Index had significant changes deleting sections, combining the active and inactive list of voters into one list that could be separated to accommodate long lines of voters and the Roster Index was out in a binders. Other changes included creating opening and closing checklists, adding two new types of voter notations to our precincts, as well as, over 100 other identified changes to the training curriculum. 10 The statement is a contradiction to the statement on page 55, “Santa Clara’s debriefing document had bulleted suggestions for improvement, such as “need more training on procedures‑role playing” and “more (polling place) closing instruction.” These changes are “material” identified by election officer feedback and Field Inspector, Election Officer and Training debriefings. 1 Audit Issue on Page 56, Paragraph 2: “Although all eight counties we visited indicated that they can receive complaint calls from voters or poll workers on election day, not all counties take steps to document the complaints that come in and how these complaints were ultimately resolve. Only three of the eight counties we visited were able to provide us with a complaint log detailing calls from poll workers, voters, or both during the February or June 2008 election.” . 6 Santa Clara County Response: The ROV provided the auditors with all Voter Feedback forms and summaries completed for the February and June 2008 elections. They contained detailed information and are the logs. The February materials were given to the auditors in May. 1 Audit Issue on Page 57, Paragraph 1: “Santa Clara County’s election services coordinator told us that summarizing complaints would be labor intensive, which would take time away from the county’s existing resources,” after stating in the proceeding sentences, “Some county officials saw limited value in recording incoming complaint information and how it was resolved.” 3 California State Auditor Report 2008-106 89 September 2008 Santa Clara County Response: We disagree with the assertion, accuracy of the quote attributed to us, and 11 context with which the statement was used. This context is clearly not Santa Clara County’s belief or correct assertion of its value. This statement is incorrectly contextually and factually quoted from the email of 8/13/08 from which we stated, “When the county receives a call, a staff member answering the phone addresses the issue and tries to resolve it immediately, if possible. If the issue cannot be handled immediately, it is recorded on a Voter Feedback form and is given to the appropriate manager who can address the concern and respond to the question or complaint. Complaints are handled as they come into the office, and an employee at the Registrar’s office responds to each complaint by resolving the complaint and then calling the voter to inform him/her that the complaint was received and handled. Documentation does not exist for all complaints in necessarily one data base, and complaints with documentation are not compiled in a database or other single document, because entering that data would be labor intensive, which would take time away from the county’s existing resources.” Santa Clara County has previously provided logs and information on the processing of emails through 8 our website. The County has emailed logs from the February and June Elections and examples of the weekly summary reports. See attached “Santa Clara Log Public Feedback Report” and EO and 299‑POLL 5 summary files.. 4 90 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 91 September 2008 Comments CAlIfORnIA STATe AudITOR’S COMMenTS On The ReSPOnSe fROM SAnTA ClARA COunTy To provide clarity and perspective, we are commenting on the response to our audit report from Santa Clara County. The numbers below correspond to the numbers we have placed in the margin of the county’s response. While preparing our draft report for publication, some wording 1 changed. In addition, page numbers shifted and therefore the page numbers the county cites may not correspond to the page numbers in our final report. Santa Clara County misrepresents the facts when it claims we 2 asked for data that we could manipulate. We repeatedly tried to work with the training data the county provided. As we state on page 36 of the report, the county provided two separate training lists for the February 2008 election; however, we noted inconsistencies between these reports and the county was unable to explain the differences. Further, had the county told us upfront that its database does not generate complete training reports, as it now states in its response, we would have reached the same conclusion much sooner. For the same reasons stated in comment #2, we were unable to 3 corroborate Santa Clara’s assertion that it ensures all inspectors attend training or they do not work. Santa Clara County describes its process for ensuring that all 4 inspectors are trained, discussing its pre‑printed class roster and other materials. Our audit methodology was to determine whether the county had a listing of trained inspectors, and if so, to verify the accuracy and completeness of such a list. While we appreciate the county’s description of its process, we found that the county’s training list was not sufficiently reliable to demonstrate that all inspectors attend training. Santa Clara County references various attachments it included with 5 its response to the audit that it provided on September 5, 2008. It chose not to provide these documents to us while we were conducting our fieldwork in May 2008. We have not included these attachments in this report; however, Santa Clara County’s attachments are available for the public to review at the California State Auditor’s Office. 92 California State Auditor Report 2008-106 September 2008 6 Santa Clara County states that it disagrees with the report’s conclusion that it could not provide us with a complaint log showing how the county responded to complaints. Although we acknowledge that its Voter Feedback forms provide some evidence of its efforts to record complaints, we do not consider them to constitute a complaint log. Further, these forms contained handwritten information making it unclear as to how the county resolved the complaints it received. 7 Santa Clara County disagrees with our conclusion, as reflected in two of the columns in Table 3 on page 37 of our report, that it could not provide documentary evidence indicating that its instructors or county staff evaluate poll worker training classes. We agree that the documentation the county refers to in its response was sufficient to support a positive rating in Table 3 under the heading “Post Election Debriefings By County Elections Staff.” However, the nature and quality of this documentation did not support a positive rating in either of the columns titled, “Evaluations Of Training By Instructors Teaching The Class,” and “Evaluations of Training By County Elections Staff.” Therefore, we stand behind the data presented in Table 3. 8 Santa Clara County disagrees with our conclusion, as reflected in two of the columns in Table 3 on page 37 of our report, that it could not provide documentary evidence that it had summarized data on poll worker questions and voter complaints. We agree that in May, the county provided us with documentation of the phone call volume it handled for the February 2008 election; however, the data provided did not indicate what the calls were about and did not indicate whether they were complaint calls. The e‑mail complaints the county refers to in its response were not provided to us in May; rather, they were provided to us on September 4, 2008, the day before its response to the audit was due. As a result, we have not considered this data in our audit, but it is available for review at the State Auditor’s Office. Finally, the county is correct that it provided a series of documents for our review. However, these documents were not summaries of complaints. The field inspector summary reports the county refers to are the written reports of three individuals who discuss their observations during the February 2008 election after visiting various polling places. In our view, the reports from these three individuals, referred to as field inspectors, do not constitute summarized data on poll worker questions. 9 Santa Clara County challenges our conclusion that we could not identify substantial changes in its poll worker training program between the February and June 2008 elections. We stand by our conclusion. Our methodology was to compare the powerpoint presentations instructors used and the hand‑out materials provided California State Auditor Report 2008-106 93 September 2008 to poll workers for those two elections, looking for changes that were obvious to us as outside observers. Although the materials we reviewed discussed items such as the roster index and referenced role‑playing exercises, we could not identify substantial changes between these two training programs. Santa Clara County has apparently misread our report and attempts 10 to raise a contradiction that does not exist. In particular, we acknowledge on page 41 of the report that its debriefing document included bulleted suggestions for improvement. However, from our review of its February and June 2008 training materials, these suggestions apparently were not implemented because the June 2008 training materials were substantially similar to those used in February. Thus, we see no contradiction and we stand behind our conclusion. We recognize that Santa Clara was able to provide some 11 documentation of its complaints in the form of documents such as its public feedback forms. However, as we state in our comment #6, these documents contained handwritten information making it unclear as to how the county resolved the complaints. We have added wording to our report to indicate that Santa Clara County was able to provide documentation of its complaint records; however, this does not change the original meaning of our report. The fact remains that Santa Clara County was unable to provide us with summarized data on the complaints it received as we discuss in our comment #8. 94 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 95 September 2008 (Agency response provided as text only.) Solano County Registrar of Voters 675 Texas Street, Suite 2600 Fairfield, CA 94533 September 5, 2008 Elaine M. Howle, State Auditor* Bureau of State Audits 555 Capitol Mall, Suite 300 Sacramento, CA 95814 Dear Ms. Howle: This is in response to your August 29, 2008 letter and the redacted draft report of the audit your staff conducted entitled: “County Poll Workers: The Secretary of State’s Office Has Developed Statewide Guidelines, But County Training Programs Need Some Improvement.” General Comments Poll worker training constantly changes from one election to the next. This is due to our belief in continual improvement, our county’s emphasis on customer service, the particular demands of a given election, and lessons learned from prior elections – including lessons from other counties. To a certain extent it is also driven by budget constraints; while there are many training tools in which we would like to invest—online training, interactive training, informal drop‑in training—limited funds require prioritizing poll worker recruitment and training goals. Independent of the BSA audit, we planned for some fairly significant changes to poll worker recruitment and training for November, 2008, and subsequent elections. We also are piloting a call center program for technical issues and Solano County is implementing a 3‑1‑1 “information center” which will track public calls before and during elections. We have redirected a large portion of scarce support funds from technical extra‑help staff to increased poll worker stipends (Stipend Schedule attached). We also have increased the stipend for attending training and provided separate training on equipment which is available to both inspectors and clerks. Along with this is a change in our philosophy of precinct board responsibilities: Where we in the past placed almost all board responsibilities on the inspector, we are giving clerks greater roles in the success of the board and, if the board successfully completes an array of assigned tasks, the entire board will receive a bonus. Our technical call center is an adjunct to our inventory management program, Asset Shadow. While still in the pilot or beta phase, we will use it in November to receive and monitor voting equipment calls from precinct workers. The operator receiving the call creates a trouble ticket and assigns it to field staff. Once the issue is resolved, the ticket is closed. Following the close of election day, we expect to have better reports of issues and resolutions than heretofore. * California State Auditor’s comments begin on page 103. 1 96 California State Auditor Report 2008-106 September 2008 Far beyond the financial means of Solano County Registrar of Voters (ROV) alone, the County has implemented a countywide 3‑1‑1 information program. Providing live operators daily from 6 a.m. to 10 p.m., these operators answer questions and create call records. When they are not able to provide an answer, the operators create a customer service request emailed immediately to ROV staff. Situations requiring more immediate ROV responses are forwarded immediately to ROV staff. Reports of calls, responses, and resolutions will be available. We note that we were one of the four counties who did not consider the Secretary of State Poll Worker Training Guidelines as having the force of law. Staff have collectively reviewed them multiple times and, 1 as the report notes, generally incorporated them into our training. We embrace them, as is noted in the auditors’ report, as suggestions rather than fiats. We further note that the SoS website identifies the electronic filename of the guidelines as “final draft” and, in the Word version, is presented with additions and 1 changes visible. Even a casual reader would find this less than definitive. The forward of the document contains internal inconsistencies as to the purpose of the document: “The following guidelines are adopted pursuant to the requirements of SB 610, and reflect the work of the Task Force. The subject of poll worker training is not so much a science as it is an art. These guidelines are intended to provide a starting point for county poll worker training programs. These guidelines will be adapted, improved and supplemented in the future as lessons are learned from field experience and voting systems change in this constantly evolving field. These guidelines are not intended to take the place of county poll worker training materials or resources. They are meant to establish a minimum set of requirements which poll worker training sessions and materials must meet and to set a standard by which local programs should be measured.” The guidelines reflect the work of a volunteer Task Force. There is no mention of public hearings, the formal 2 acceptance process of the Secretary of State’s Office, supporting directives or any other trail of documents typical in the promulgation of a regulation or law. The forward of the document refers to poll worker training as being an art and that the guidelines would need to be adapted, improved, and supplemented. This has not occurred as the document, as “published,” does not even reflect the directives and guidelines issued by the Secretary with regard to voting systems resulting from the Top to Bottom Review. While the guidelines offer reasonable suggestions, taken as a whole, it is doubtful that all the suggestions could be incorporated into a training program that would meet reasonable training session times and be affordable. 2 1 It is therefore the opinion of our Office that the guidelines are simply just that, guidelines. To be binding on counties, the guidelines would require conversion into statutes or regulations and a state‑mandated cost would ensue. A recent published decision by the State Court of Appeal, Fourth Appellate District, (D052744) confirms this view. On August 29, 2008, the court nullified a Secretary of State directive for its failure to be enacted through the regulatory process prescribed by the Administrative Procedures Act. 2 California State Auditor Report 2008-106 97 September 2008 Further, it is questionable that any governmental or oversight body has the authority to “audit” counties’ 3 1 2 conformance or lack thereof to a concept or guideline for which no mandate has been established and for which the process has been described as “not so much a science as it is an art” within the guideline itself. The tone of the report is negative and accusatory and in our opinion does not reflect the reality of the elections environment. The auditors seemed to be overly concerned with voter complaints (whether 4 justified or not) as opposed to the successful conduct of the election process and the official canvass of votes. The conduct of elections relies heavily on tens of thousands of poll workers who are essentially volunteers performing a very complex task either for the very first time or possible once or twice a year. In addition to the normal complexities of a primary election, the bifurcated February 2008 primary election generated widespread confusion about the nature of the modified primary system in general. The process was simply not intuitive for voters or poll workers and an over emphasis on an specified number of voter complaints does not do justice to our state’s poll workers or County election offices especially given that the report actually shows that counties generally comply with the guidelines. The report often refers to “state law.” It would be useful for the auditors to have cited particular Elections 5 Codes or other applicable state codes to provide reference and context to their remarks. Our response comments on the report overall, specific references where they affect Solano County, and the recommendations. The Report Generally, the report suffers from isolated exposure to our training (11 percent of our training sessions) 6 and the unwillingness of the auditors to adequately analyze materials given them, and then leaps to universal conclusions and contradictory recommendations. At the very least, the report should begin with a disclaimer that auditors’ observations were very limited, that there was insufficient time to adequately review all materials submitted, the findings should not be considered absolute for any county surveyed, and the recommendations are accordingly general and may not be universally appropriate. Further, an underlying 7 but unspoken tenet of the recommendations is that more money should be spent on poll worker training; this fundamental truth should be made loud and clear. It is entirely foreseeable that the Legislature will selectively embrace the report’s findings and recommendations and attempt to inflict unfunded poll worker training mandates on counties.. The auditors’ apparently had a mindset that could only deal with information presented in a linear, heavily 8 documented fashion: That following the February, 2008, presidential primary, counties would undertake a thorough analysis of poll workers’ performance, poll worker evaluations and comments, and voters’ comments (universally referred to as “complaints” in the report); that analysis would result in a detailed report; and all training materials for the June, 2008, primary would be carefully adjusted in concert with the findings. Since we do not have sufficient staff for such an exercise and there was too little time between elections, the auditors say information we provided is not reliable; this is not the case. It also dismisses the 8 validity of the decades of election experience held by Solano County ROV staff and our ability to identify poll worker performance issues, triage those that most adversely affect successful poll worker performance, and develop corrective training and management measures. 3 98 California State Auditor Report 2008-106 September 2008 While we were promised and given an “exit interview” where we would have the opportunity to review the report before it went into print, the interview was too short and useful mainly to confirm that auditors did 9 not fully comprehend what we said. With BSA staff paraphrasing report text, there was little opportunity to meaningfully address the many significant differences and contextual representations in the report. In 10 commenting on the redacted report, we find ourselves limited in effectively responding to the redundancy in the report and the BSA caveat that references to page numbers are likely to change and text may have changed since the hardcopy was delivered. Observations 11 1. The last paragraph on page 6 says, “However, not all counties (handwritten note: includes Solano) were able to provide data that demonstrates they trained all inspectors prior to the February 2008 election.” This is repeated on page 48 where the report says, “Solano county was unable to provide us with a listing of poll workers who received training prior to the February 2008 election. Instead, the county only provided the sign‑in sheets used at various trainings. Since Solano County did not provide a training attendance list, and we could not verify that it had provided sign‑in sheets for its poll worker trainings, we did not perform further anaylsis.” For someone not familiar with the poll worker training and payroll possibilities in Solano County, determining who attended training is not as simple as checking names off a list titled “Training 12 Attendance List.” Despite repeated attempts to explain our processes of checks and balances, the auditors either did not understand or did not accept our procedures. All inspectors were trained prior to the February 2008 election. They could not receive precinct ballots and other essential supplies without 12 being trained. We offered to provide additional documentation which they said was not permitted because of unspecified restrictions. The auditors could have also followed up with all, or a sampling, of our poll workers to verify that they attended training. 11 Page 34 of the report says, “However, not all counties (redacted) were not able to (handwritten includes Solano) provide data that demonstrated they trained all inspectors prior to the February 2008 election. As a result, the counties cannot be certain to what extent these workers who supervise polling places have the knowledge to efficiently administer elections.” 12 This statement is incorrect on several points. First, we did provide or offered to provide the auditors with data that demonstrated all inspectors were trained; they chose to not accept it. Second, whether we provided auditors data demonstrating that all inspectors had been trained has no bearing on our certainty of inspectors’ capabilities to manage their precinct board. Third, even when inspectors take advantage of all training opportunities and have experience over many elections, there is always the possibility for misapplying procedures or election law. The auditors’ implicitly acknowledge this in their 13 recommendations for multiple training opportunities.. 11 2. The report says on page 7, “Moreover, the election officials from eight counties we visited told us they use a variety of sources for collecting information for identifying needed improvements in their poll worker training programs. Sources the counties told us they used included post‑training feedback from poll workers, comments from instructors, post‑election debriefing reports, analyses of voter complaints, and reviews of questions from poll workers on election day. However, seven of the eight counties were able to provide documentation of the information they collected.” This is supplemented by Table 3 and repeated on pages 34 and 50. 4 California State Auditor Report 2008-106 99 September 2008 The auditors relied on documentation‑based methods of determining the need for changes to poll 14 worker training. However, there is no legal, or even generally accepted, criteria within the elections community for measuring performance on this process. We submit that our principle means for analyzing poll worker performance—the ease and accuracy of reconciling the final canvass—is more meaningful than compiling anecdotal comments. That our analysis doesn’t result in a report leaves it 14 out of the auditors’ realm of comprehension. However, our approach is most closely tied to the integrity and success of the election. When poll workers perform their jobs well, the final canvass is quicker, easier to understand and explain, and more accurate. 3. Page 7 of the report says, “Under state law, voters have the right to ask poll workers and elections 11 officials questions and complaints about election procedures and to receive an answer or be directed to an appropriate elections official for an answer.” The wording and repeated references to “complaints” here and throughout the audit process and report 4 is troubling. The Elections Code makes clear, and we embrace wholeheartedly, the election process is open to all. EC § 2300, the Voter Bill of Rights, specifically addresses access: (9) (A) You have the right to ask questions about election procedures and observe the election process. (B) You have the right to ask questions of the precinct board and elections officials regarding election procedures and to receive an answer or be directed to the appropriate official for an answer. However, if persistent questioning disrupts the execution of their duties, the board or election officials may discontinue responding to questions. (10) You have the right to report any illegal or fraudulent activity to a local elections official or to the Secretary of State’s office. Other Elections Code sections similarly reinforce the openness of elections (e.g., §§ 15620 – 15634, final canvass; § 15104, vote by mail processing; § 14403, closing of the polls). The “complaints” refrain is picked up again on page 35 (and 56–57), saying, “And while all eight counties 11 told us they receive complaint calls from voters or poll workers on election day, only three (handwritten does not include Solano) were able to provide us with a complaint log detailing those calls for either the February 2008 or June 2008 elections.” While we lack the documentation required of the auditors to be believable, our experience is that we very few complaints. The vast majority of our calls from voters have to do with polling place location, registration, and vote by mail status. We note further that the Elections Code does not require that we keep detailed logs of complaints, questions, 15 our responses or resolutions, or for any other reason. As stated earlier, Solano County has embarked on an ambitious 3‑1‑1 information program that may result in the kinds of reports the auditors’ desire. 4. The auditors on page 37 correctly say that in the one inspector class they observed that operation of the AutoMARK disabled voter assistive device was not covered because training ran overtime although the instructor offered to stay for individual instruction. This points to a contradiction in 16 5 100 California State Auditor Report 2008-106 September 2008 recommendations addressed below, where auditors recommend multi‑tiered training for experienced vice new poll workers. If everything in the guidelines is required, why offer different training? In the training auditors observed, almost all of the attendees were experienced inspectors and had been through training in October 2007. Inspector training and use of equipment is also supplemented by trained precinct rovers and technicians in the field. Recommendations (Note: recommendations are repeated in the report) 11 1. The auditors’ first recommendation, on page 9, says, “To ensure that poll worker training programs comply with the Secretary of State’s guidance, county elections officials should review the content of their programs, ensuring that they fully cover topics such as voter complaint procedures, preventing voter intimidation (emphasis added), and issues pertaining to a culturally diverse electorate.” Voters have a right to cast a ballot “free from intimidation” EC § 2300(a)(4). Voter intimidation is defined in EC §§ 18540–18548. Poll workers generally are able to only affect intimidation opportunities by restricting electioneering within 100 feet of the polling place. 2. The auditors offer three recommendations “To improve poll workers’ willingness to attend training and ability to retain lessons learned, county officials should consider implementing the following practices: • Maximize the number of poll worker training sessions while also offering the training at multiple locations with different start times to better accommodate poll workers’ other time commitments. Also, providing condensed training to “experienced” poll workers may entice greater attendance, while more extended training can be reserved for “new” poll workers. Comments: As noted earlier, if training must cover everything in the guidelines plus 16 county‑particular topics, having sessions for “new” vice “experienced” poll workers creates an added scheduling burden in the training calendar. There are also advantages to having the “voice of experience” from poll workers who can attest to our training points. • Offer poll workers an opportunity to reinforce what was learned in class through the use of online supplemental training material. Such an online program might include practice quizzes on election‑day procedures, examples of the election materials to be used, and reference material provided at training. County election officials might also consider providing podcasts that emphasize critical aspects of poll worker training. • Provide poll workers with additional opportunities to remember what they learned and get hands‑on experience with using election‑day supplies and voting equipment through optional workshops. County elections officials might consider providing these workshops in the days immediately before an election to maximize poll worker confidence and retention of information. Comments: As noted earlier, the auditors’ recommendations contain an underlying commitment of funds which are in increasingly short supply. The prospect of serious short‑falls in the state’s budget and the prospect of a 2009 statewide special election will likely adversely affect county budgets. Counties using online training have increasing support costs and note that online training is, at best, an adjunct to live, in‑person training. The auditors also fail to mention a much‑proven means of getting poll workers to training (which we repeatedly offered): pay the poll workers more money to be a poll worker and attend training. 6 California State Auditor Report 2008-106 101 September 2008 3. The auditors offer two recommendations “To better ensure that county election officials provide knowledgeable poll workers to serve voters, counties should take steps to ensure that all inspectors receive training. Steps that counties might take to achieve this goal include: • Compiling accurate lists of inspectors that have attended training, while informing inspectors that did not go through training that they cannot serve as inspectors. • Recruit reserve poll workers that have gone through inspector training, who can be deployed, as necessary, to polling places where assigned instructors (sic) had not been trained. Comments: We have complete and accurate lists of trained inspectors. That the lists are not in the 17 format preferred by the auditors does not make them less valuable to us. We try to have a cadre of trained poll workers who can fill in as necessary. 4. Finally, under Summary, the auditors recommend “To better ensure poll worker training programs are effectively evaluated and needed improvements identified, county elections officials should consider taking steps to track voter complaints and poll worker questions that are received during an election, evaluating whether such comments suggest ways to improve their training programs and implementing those improvements. Comments: The auditors fail to suggest that the analysis result in a written report targeting specific changes to training topics. Regardless, the auditors again rely too heavily on “complaints” as indicators 4 of a need for changes to training. While there may be value to compiling comments from voters and poll workers (which we did for the June 2008 election, although the auditors failed to mention it in the 18 report) analysis of precinct board errors and omissions provides a very direct indicator of training needs. Providing a board with a bonus for very good performance and denying a bonus from those who fail stated performance goals reinforces targeted training topics. In summary, we approached this audit with the spirit of cooperation and the hope that BSA would follow through on their opening promises of a free exchange of information and a product the BSA and counties would be proud of and the Legislature would find useful. We submit our response to the report very disappointed that the BSA selectively, and apparently arbitrarily since we were never informed of what constituted valid data, chose what information to use. The ultimate product is more a review of county 19 record‑keeping practices (in an area where no record‑keeping requirements exist) than a meaningful report of how well poll workers are trained. Sincerely, (Original not signed) Ira Rosenthal Chief Information Officer Registrar of Voters Lindsey McWilliams Assistant Registrar of Voters 7 102 California State Auditor Report 2008-106 September 2008 Blank page inserted for reproduction purposes only. California State Auditor Report 2008-106 103 September 2008 Comments CAlIfORnIA STATe AudITOR’S COMMenTS On The ReSPOnSe fROM SOlAnO COunTy To provide clarity and perspective, we are commenting on the response to our audit report from Solano County. The numbers below correspond to the numbers we have placed in the margin of the county’s response. Our report accurately describes state law. As we state on page 1 of 1 the report, the purpose of the office’s training guidelines was to establish a minimum set of requirements that training sessions and materials developed by the counties must meet and to set a standard against which county programs for poll workers should be measured. In fact, the Election Code, Section 12309(a) states that a county’s training program for inspectors shall conform [emphasis added] to the uniform standards adopted by the Secretary of State. Further, as we state on page 16 of the report, the office issued its final training guidelines on April 19, 2006. Finally, the training guidelines we reviewed on the office’s Web site are not in draft form. Solano County claims that the lack of public hearings or formal 2 acceptance by the office supports its contention that the training guidelines are not requirements. According to our legal counsel, it appears that the Legislature did not intend for the Secretary of State to go through the typical rulemaking process required when state agencies adopt regulations. Instead, the legislation calls for the establishment of a task force of experts who were required to make their recommendations available for public review and comment prior to their submission to the Secretary of State and the Legislature. Thus, while Solano suggests that the uniform standards were not publicly vetted, the task force was required by law to do so. Nonetheless, if a court of law found that the uniform standards should have gone through the typical rulemaking process, the Secretary of State could seek to have the uniform standards approved under that process. Further, the office has formally adopted its training guidelines and notified counties as discussed in our comment #1. In addition, Solano claims that if Section 12309(a) requires a county’s training program to conform to the uniform standards, then it would be a state mandated local program and the state would be required to reimburse counties for their costs. Our legal counsel has advised us that when a bill creates a state‑mandated local program, Legislative Counsel is required by law to indicate that within the bill. Legislative Counsel did not identify this legislation as creating a new state mandate. According to our legal counsel, the likely reason is that counties were already required to provide training to inspectors, thus the legislation did 104 California State Auditor Report 2008-106 September 2008 not require counties to create a new program or provide a higher level of service and therefore counties would not be incurring new or additional costs. 3 Solano County questions our authority to audit it for conformance to the office’s training guidelines. Our authority to audit Solano County can be found in the California Government Code, Section 8546.1(b) that states, in part, “[t]he State Auditor shall conduct any audit of a state or local governmental agency or any other publicly created entity that is requested by the Joint Legislative Audit Committee.” On March 12, 2008, the Joint Legislative Audit Committee approved the audit request and directed the State Auditor’s Office to conduct the audit. 4 We are puzzled by the county’s comment. Our audit report covers many aspects of poll worker training and does not focus exclusively on how counties respond to complaints. Rather, the report includes discussion regarding training guidelines, monitoring of elections, and poll worker training as well as counties’ handling of complaints. 5 Solano County objects to our use of the term “state law” in our audit report. To make our reports more reader friendly, we do not always provide specific statutory references. However, we believe we clearly specify what the law requires. However, the public workpapers that support our report, including all legal citations, are available for review at our office. 6 Solano County challenges our report’s conclusions by claiming that we had isolated exposure to its training. We assume that Solano’s comment is meant to infer that we did not observe more of its training sessions. The conclusions in our report are based on multiple procedures and different audit evidence obtained. In addition to our direct observation of a training session, we reviewed documents provided to poll workers during training. We also obtained testimonial evidence from county elections officials when appropriate. We believe this body of work supports our conclusions. 7 Solano County has misread and misunderstands the point of our recommendations. In particular, we are not suggesting that the counties spend more money. Instead, the recommendations beginning on page 47 of the report are offered to help all counties improve the poll worker training programs that they have already been required to conduct beginning with the November 2008 election. 8 Solano County’s comments suggest that it was frustrated with our focus on obtaining documentation. Audit standards require us to obtain sufficient and appropriate evidence to support our California State Auditor Report 2008-106 105 September 2008 conclusions. Audit evidence can take the form of an auditor’s direct observation, documentary evidence, or testimonial evidence. In our judgment, relying on the county’s testimonial evidence for certain audit conclusions, without the benefit of corroborating documentary evidence, would not comply with audit standards and would not provide a sufficient basis for our conclusions. Solano County incorrectly claims it had limited opportunity to 9 meaningfully address the issues in the audit report. After reading portions of the report to county officials during our exit conference, we consistently asked for their perspective on what they heard. Further, we repeatedly asked Solano if they had questions about our conclusions. Finally, we provided Solano with five business days to review a draft of the report so that it could provide its response to the audit. Government Code, Section 8545(b) prohibits the State Auditor 10 from releasing to the public any information about an audit that is not yet completed. Thus, when there are multiple counties involved in an audit, as was the case here, we can only share what we find at a particular county with that county while the audit is ongoing. While preparing our draft report for publication, some wording 11 changed. In addition, page numbers shifted and therefore the page numbers the county cites may not correspond to the page numbers in our final report. Solano County’s response reinforces a conclusion in our report 12 that the county does not have an attendance list indicating which poll workers attended training. Our audit methodology was to determine whether the county had a listing of trained inspectors, and if so, to verify the accuracy and completeness of such a list. As we state on page 36 of the report, Solano County could only provide us with sign‑in sheets from its training sessions. Solano County offered additional documentation in the form of receipts for supplies, as it mentions in its response, but we found these receipts lacked dates and could not be matched with the dates the county provided the training. As such, we could not gain assurance that the receipts for supplies were linked to the training sessions as Solano County asserts. We agree that experienced poll workers may forget what was 13 learned, especially when training takes place weeks before the election. For this reason, we saw that some counties had a practice of offering optional workshops in the days leading up to the election. Providing this type of optional training seemed to be a notable practice to help maximize knowledge retention among poll 106 California State Auditor Report 2008-106 September 2008 workers as we describe on page 33 of our report. However, we fail to see how this point, in any way, diminishes the importance of ensuring that all inspectors attend training. 14 Solano County is challenging our decision to require documentation in order to receive a positive rating in Table 3 on page 37 of the report. As we clearly note in Table 3, we required counties to provide documentary evidence of their data collection practices in order to receive a positive rating. Audit standards require us to obtain sufficient and appropriate evidence to support our conclusions. Audit evidence can take the form of an auditor’s direct observation, documentary evidence, or testimonial evidence. In our judgment, relying on the county’s testimonial evidence to support our conclusions, without the benefit of corroborating documentary evidence, would not comply with audit standards and would not provide a sufficient basis for our conclusions. Finally, our report does not conclude that counties fail to employ the methods shown in Table 3. Instead, our point is that we did not always see evidence of those methods in practice. 15 We are aware that the Election Code does not require counties to keep detailed logs of complaints, questions, or the county’s responses and resolutions. Nevertheless, we refer the county to our comment #14 for our explanation of why we required counties to provide us with documentary evidence of their practices. 16 Solano County incorrectly claims that our recommendations are inconsistent with the report’s text. Counties can have shortened training sessions for experienced poll workers, or longer sessions for new poll workers, as long as the material presented to inspectors conforms with the office’s poll worker training guidelines. 17 Solano County asserts that it has complete and accurate lists of trained inspectors. If this were true, it should have provided them to us during the audit. Further, as described in our comment #12, the documentation Solano County provided was not sufficient to determine whether all inspectors were trained. 18 Solano County claims that we failed to mention that it compiled comments from voters and poll workers for the June 2008 election. We would have made such a statement if the county was able to provide documentary evidence demonstrating that it had done so. Instead, the county only provided a blank form that it asserted was going to be used to collect such feedback. Nevertheless, we are pleased that Solano sees value in collecting such data. 19 Solano claims that our report is a review of county record‑keeping practices. We disagree. The scope of our audit and the methodology we used are clearly described on pages 11 through 14 of our report. California State Auditor Report 2008-106 107 September 2008 cc: Members of the Legislature Office of the Lieutenant Governor Milton Marks Commission on California State Government Organization and Economy Department of Finance Attorney General State Controller State Treasurer Legislative Analyst Senate Office of Research California Research Bureau Capitol Press