This is the accessible text file for GAO report number GAO-05-997 entitled 'Elections: Views of Selected Local Election Officials on Managing Voter Registration and Ensuring Eligible Citizens Can Vote' which was released on October 27, 2005. This text file was formatted by the U.S. Government Accountability Office (GAO) to be accessible to users with visual impairments, as part of a longer term project to improve GAO products' accessibility. Every attempt has been made to maintain the structural and data integrity of the original printed product. Accessibility features, such as text descriptions of tables, consecutively numbered footnotes placed at the end of the file, and the text of agency comment letters, are provided but may not exactly duplicate the presentation or format of the printed version. The portable document format (PDF) file is an exact electronic replica of the printed version. We welcome your feedback. Please E-mail your comments regarding the contents or accessibility features of this document to Webmaster@gao.gov. This is a work of the U.S. government and is not subject to copyright protection in the United States. It may be reproduced and distributed in its entirety without further permission from GAO. Because this work may contain copyrighted images or other material, permission from the copyright holder may be necessary if you wish to reproduce this material separately. Report to Congressional Requesters: United States Government Accountability Office: GAO: September 2005: Elections: Views of Selected Local Election Officials on Managing Voter Registration and Ensuring Eligible Citizens Can Vote: GAO-05-997: GAO Highlights: Highlights of GAO-05-997, a report to congressional requesters: Why GAO Did This Study: GAO’s past work and the work of others has shown that challenges processing voter registration applications and maintaining voter registration lists can result in individuals arriving at polls on Election Day to find they were not listed as registered. GAO surveyed local election officials in 14 jurisdictions in 7 states (AZ, CA, MI, NY, TX, VA, and WI) to obtain their views on managing voter registration for the 2004 election. GAO selected the 7 states considering characteristics relevant to voter registration, such as whether a statewide voter registration list existed prior to the enactment of the Help America Vote Act (HAVA) of 2002. Locations were selected within each state to represent one small and one large election jurisdiction. This report discusses election officials’ characterization of (1) challenges receiving voter registration applications, including checking them for completeness; (2) removing voters’ names from voter registration lists and ensuring that names were not inadvertently removed; and (3) implementing HAVA’s provisional voting and identification requirements. HAVA, in part, requires that states offer provisional ballots to voters not listed as registered who declare eligibility and first-time voters who registered by mail after January 1, 2003, and could not provide identification. GAO offered election officials the opportunity to verify the accuracy of their responses used to prepare this report. What GAO Found: Local election officials representing all but 1 of the 14 jurisdictions GAO surveyed after the November 2004 election reported facing some challenges processing voter registration applications and took steps to address them. Processing applications received from voter registration drives sponsored by non-governmental organizations posed a challenge to election officials in 12 of the 14 jurisdictions, while half of the officials reported challenges receiving applications from other external sources, such as motor vehicle agencies. Challenges occurred in processing these applications for reasons such as incomplete or inaccurate information on voter registration applications. Half of the officials reported that their offices faced challenges checking applications for completeness, accuracy, or duplicates, citing, among other things, insufficient staffing to check the applications. Steps taken by election officials to address these and other challenges included hiring additional staff to handle the volume of applications received and contacting applicants to get correct information. All but 1 of the 14 election officials reported that, using various sources of information, they removed names from voter registration lists during 2004 if, for example, voters had moved, were deceased, or were ineligible due to a felony conviction. To help ensure names of eligible voters were not inadvertently removed from voter registration lists, officials reported contacting voters to confirm removal, matched voters’ identifying information (such as name and address) with address changes provided by the U.S. Postal Service, and matched voter registration records with felony or death records. GAO reported in June 2005 about problems officials in these same jurisdictions experienced verifying voter information with death or felony information from existing data sources. GAO’s survey showed that all 14 election jurisdictions permitted citizens to cast provisional ballots during the November 2004 election. HAVA gives states discretion to implement provisional voting based on state voter eligibility requirements. According to the election officials surveyed, about 423,000 provisional ballots were cast in 13 of the 14 jurisdictions, and 70 percent of those votes were counted. Also, 8 of the 14 jurisdictions reported challenges implementing provisional voting, in part, because some poll workers were not familiar with provisional voting or staff did not have sufficient time to process provisional ballots. To address these challenges, election officials in these jurisdictions said they hired extra staff or provided training to poll workers. What GAO Recommends: www.gao.gov/cgi-bin/getrpt?GAO-05-997. To view the full product, including the scope and methodology, click on the link above. For more information, contact William Jenkins at (202) 512-8777 or jenkinswo@gao.gov. [End of section] Contents: Letter: Results in Brief: Background: Selected Jurisdictions Reported Challenges Processing Voter Registration Applications and Various Steps They Took to Address Them: Jurisdictions Reported They Removed Voter Names from Registration Lists for Various Reasons, but Took Steps to Ensure That Eligible Voters' Names Were Not Inadvertently Removed: All Jurisdictions Reported Enabling Provisional Votes to Be Cast but Not All Provisional Ballots Met States' Criteria for Whether These Votes Should Count: Appendix I: Scope and Methodology: Appendix II: Survey of Local Election Officials: Appendix III: Reported Experiences of Selected Local Election Offices Receiving Voter Registration Applications: Appendix IV: Reported Experiences of Selected Local Election Offices Checking Voter Registration Applications: Appendix V: Reported Experiences of Selected Local Election Offices Entering Information into Voter Registration Lists: Appendix VI: Reported Experiences of Selected Local Election Offices Removing Names from Voter Registration Lists: Appendix VII: Reported Experiences of Selected Local Election Offices Implementing Provisional Voting: Appendix VIII: Reported Experiences of Selected Motor Vehicle Agency Offices with Voter Registration Applications: Experiences of Selected State MVA Offices: Experiences of Selected Local MVA Offices: Appendix IX: Survey of Motor Vehicle Agency Officials: Appendix X: GAO Contact and Staff Acknowledgments: Tables: Table 1: Election Officials' Responses to Questions about Reasons Why Voters' Names Were Removed from Voter Registration Lists in 14 Jurisdictions during 2004: Table 2: Number and Percentage of Provisional Ballots Election Officials Reported Cast and Counted in the November 2004 Election in 14 Jurisdictions: Table 3: Number of Provisional Ballots Cast as a Percentage of Total Ballots Cast for President in the November 2004 Election in 14 Jurisdictions: Table 4: State Selection Factors: Figures: Figure 1: Examples of Opportunities Citizens Have to Apply to Register to Vote: Figure 2: Sample of Michigan's Voter Registration Application: Figure 3: Sample of Texas' Voter Registration Application: Figure 4: Extent of Challenges Reported by 14 Election Jurisdictions Receiving Applications from Organizations External to Election Offices: MVAs, Other NVRA Agencies, and Voter Registration Drives Sponsored by Non-Governmental Organizations: Figure 5: Extent of Challenges Election Officials Reported Experiencing When Checking Applications for Completeness, Accuracy, or Duplicates: Figure 6: Extent of Challenges Reported by Election Officials Entering Voter Information into Voter Registration Lists and Scanning Signatures of New Voters into Voter Registration Systems: Figure 7: Reasons That Provisional Ballots Were Not Counted for the November 2004 Election in Selected Jurisdictions: Figure 8: MVA Officials' Characterization of How Difficult or Easy It Was for MVA Staff to Assist Clients with Completing Voter Registration Applications, Accept Applications from Clients, and Forward Applications to Election Offices: Abbreviations: FEC: Federal Election Commission: HAVA: Help America Vote Act of 2002: MVA: motor vehicle agency: NVRA: National Voter Registration Act of 1993: USPS: U.S. Postal Service: United States Government Accountability Office: Washington, DC 20548: September 27, 2005: The Honorable Joseph I. Lieberman: Ranking Minority Member: Committee on Government Affairs and Homeland Security: United States Senate: The Honorable Henry A. Waxman: Ranking Minority Member: Committee on Government Reform: House of Representatives: The Honorable John Conyers, Jr.: Ranking Minority Member: Committee on the Judiciary: House of Representatives: After the November 2000 election, reports in the media and elsewhere have questioned the mechanics and effectiveness of voter registration and voter registration list maintenance by highlighting accounts of individuals who thought they were registered being turned away from polling places on Election Day. In managing the voter registration process and maintaining voter registration lists, state and local election officials must balance two goals--(1) minimizing the burden on eligible people of registering to vote and (2) ensuring that voter lists are accurate (limited to those eligible to vote), which includes ensuring that appropriately registered voters are not inadvertently removed from the voter registration lists. All the while, state and local election officials must try to ensure that eligible citizens have the opportunity to cast their votes and have them counted accurately in federal, state, and local elections. Over the last several years, federal legislation has been enacted to expand the opportunities for citizens to become registered to vote, improve the accuracy of voter registration lists, and ensure that eligible voters have the opportunity to cast their ballots. In 1993, the National Voter Registration Act (NVRA), commonly called the Motor Voter Act, was enacted.[Footnote 1] Among other things, NVRA expanded the opportunity of citizens to register to vote in federal elections at state motor vehicle agencies and other public organizations, such as public assistance agencies and armed forces recruiting centers. NVRA also limited the circumstances under which states could remove the names of eligible voters from registration lists for federal elections and required states to take certain steps to accurately maintain such voter registration lists by removing the names of certain types of ineligible persons. In 2002, the Help America Vote Act (HAVA) was enacted and, among other things, it requires states to implement provisional voting for elections for federal office. HAVA, in general, requires that individuals not listed as registered or whose eligibility is questioned by an election official must be notified about and permitted to cast a provisional ballot that is set aside for review by election officials at a later time so that they can determine whether the person is eligible to vote under state law.[Footnote 2] HAVA also requires that provisional ballots be provided to first-time voters who had registered to vote by mail on or after January 1, 2003, but were unable to show photo identification or another qualifying identification document when voting in person or by mail in a federal election.[Footnote 3] In addition, HAVA requires that election officials must provide access to information that permits voters to learn if their provisional ballot was counted, and, if not, why not. Our work, which began in January 2004, focuses on the efforts of local election officials in 14 jurisdictions within 7 states to manage the registration process, maintain accurate voter registration lists, and ensure that eligible citizens in those jurisdictions had the opportunity to cast ballots during the 2004 election. Specifically, for the 2004 election, we are reporting on election officials' characterization of their experiences with regard to (1) managing the voter registration process and any challenges related to receiving voter registration applications; checking them for completeness, accuracy, and duplication; and entering information into voter registration lists; (2) removing voters' names from voter registration lists and ensuring that the names of eligible voters were not inadvertently removed; and (3) implementing HAVA provisional voting and identification requirements and addressing any challenges encountered related to these requirements. We are also providing information on motor vehicle agency (MVA)[Footnote 4] officials' characterization of their experiences assisting citizens who apply to register to vote at MVA offices and forwarding voter registration applications to election offices. This information is contained in appendix VIII. In addition to this report, we have also issued, or plan to issue this year, reports on other specific election issues. In June 2005, we issued a report on the efforts of state and local election officials in seven states--the same seven states and local jurisdictions covered in this report--to ensure that voter registration lists are accurate.[Footnote 5] Later this year we plan to issue reports on (1) actions by nine states without HAVA waivers[Footnote 6] to create and maintain computerized, statewide voter registration lists as of January 1, 2004; (2) significant security and reliability concerns that have been identified for electronic voting systems; and (3) the Department of Defense's implementation of the Federal Voting Assistance Program to facilitate absentee voting by military personnel during the November 2004 election. These reports respond to congressional requests made prior to the November 2004 election. In addition, given concerns about the November 2004 election process, we are undertaking a broader, more comprehensive study of election administration and processes related to the November 2004 general election. This more comprehensive study, which we plan to issue in 2006, will address activities and challenges- -people, processes, and technology--associated with each major stage of election administration to include registration, absentee and early voting; Election Day preparation and activities; and vote counting and certification. In October 2001, we issued a similar report that focused on how elections were conducted in the United States, and the people, processes and technology that were generally associated with the preparation for and administration of elections. Among other things, the report discussed the activities and challenges associated with each stage of election administration, including voter registration.[Footnote 7] To address our objectives for this report, we analyzed information collected from elections and motor vehicle agency offices in seven states--Arizona, California, Michigan, New York, Texas, Virginia, and Wisconsin.[Footnote 8] Our selection of the seven states is geographically diverse, and took into consideration a range of voter registration-related factors and unique characteristics of the states that might affect the implementation of HAVA. Specifically, we selected states that took various approaches for administering elections--for example, Wisconsin has same-day registration, which exempts the state from the HAVA provisional voting requirement; Arizona has on-line registration; Michigan has a reputation as a model for registration practices; and some local jurisdictions administer elections at the county level and others at the city or township level. Within each of the seven states, using population data from the 2000 U.S. Census, we then selected two jurisdictions: a local jurisdiction with a large population and a local jurisdiction with a small population.[Footnote 9] Other criteria we applied to select these jurisdictions included the proximity of the locations to our site visits with state election and motor vehicle officials, suggestions by state election officials, and proximity to a local motor vehicle office. The 14 jurisdictions we selected were Gila and Maricopa Counties, Arizona; Los Angeles and Yolo Counties, California; City of Detroit and Delta Township, Michigan; New York City and Rensselaer County, New York; Bexar and Webb Counties, Texas; Albemarle and Arlington Counties, Virginia; and the cities of Franklin and Madison, Wisconsin. The selected states and jurisdictions are the same states and jurisdictions we discussed in our June 2005 report. We gathered information from state and local officials using a multi- pronged approach. First, we reviewed relevant laws, state reports, and documents related to the voter registration process in the seven states. Second, we interviewed state and local election officials in the seven states and 14 jurisdictions to obtain information on their registration processes and implementation of the HAVA requirements for provisional voting and voter identification. Third, we sent a survey to election officials in the 14 jurisdictions to gather information about their experiences with the November 2004 election. We developed our survey questions on the basis of our review of laws, reports, and documents and our interviews with state and local election officials in the seven states and 14 jurisdictions. Our survey primarily asked questions on (1) challenges, if any, in processing voter registration applications--specifically, challenges receiving voter registration applications from MVAs, other NVRA agencies, and voter registration drives sponsored by non-governmental organizations; checking voter registration applications for completeness, accuracy, and duplicates; and entering voter information into registration lists or systems; (2) whether names of registrants were removed from voter registration lists, reasons why names were removed, and steps taken to ensure that names of eligible registrants were not inadvertently removed; and (3) HAVA's provisional voting and first-time voter identification requirements--specifically, the number of provisional ballots cast and counted, including reasons why jurisdictions did not count provisional ballots if this did occur; how voters were informed about provisional voting and its outcome; and whether first-time voters who registered by mail were allowed to cast provisional ballots if they did not provide required identification. Appendix II provides a copy of the survey we sent to each jurisdiction. Finally, to provide information on MVA officials' characterization of their experiences with voter registration applications received before the November 2004 election, we sent a survey to state and local MVA officials in six of the seven states and 12 of the 14 jurisdictions.[Footnote 10] The survey primarily asked questions about the MVA offices' experiences with (1) assisting citizens with completing voter registration applications, (2) forwarding the applications to election offices, and (3) responding to individuals and state or local election officials who contacted their offices about individuals who declared they had applied to register to vote at MVA offices but their names were not on voter registration lists when they went to vote in the November 2004 election. Appendix IX provides a copy of the survey we sent to the MVA officials. We did not independently verify the accuracy or completeness of responses provided from our surveys of election officials and MVA officials. However, local election officials and state and local MVA officials we surveyed were provided the opportunity to verify the accuracy of their responses for this report and, on the basis of the comments we received, we made technical changes, where appropriate. The results of our work cannot be generalized to state and local election jurisdictions and MVA offices at the state level or nationwide. We conducted our work from January 2004 through September 2005 in accordance with generally accepted government auditing standards. Appendix I discusses our scope and methodology in greater detail. Results in Brief: Election officials representing all but one of the jurisdictions we surveyed following the November 2004 election said they faced some challenges managing the voter registration process, including (1) receiving voter registration applications; (2) checking them for completeness, accuracy, and duplication; and (3) entering information into voter registration lists; when challenges occurred, election officials reported they took various steps to address them. For example, when asked whether their staff faced challenges receiving voter registration applications from MVAs, other NVRA agencies, or voter registration drives sponsored by non-governmental organizations, election officials representing 12 of the 14 jurisdictions responded that staff faced challenges receiving voter registration applications from voter registration drives; 4 faced challenges receiving applications from MVAs; and 4 faced challenges receiving applications from other NVRA agencies. Officials responded that the reasons these challenges occurred included applications that were incomplete or inaccurate, and applications that were received too close to, or after, their state's registration deadline. Election officials reported that, to address these challenges, their jurisdictions took steps, including contacting applicants to resolve issues with incomplete or inaccurate applications or hiring staff to handle the volume of applications received near registration deadlines. In addition, officials in 7 of the 14 jurisdictions reported that their staff faced challenges checking voter registration applications for completeness, accuracy, or duplicates. According to these officials, these challenges occurred for a variety of reasons, including problems contacting individuals to obtain complete and accurate information and insufficient staffing to check the applications. They reported that, among other things, their staff addressed these challenges by sending letters or calling applicants to obtain correct information. Finally, 6 of the 14 election officials reported that their staff faced challenges entering or scanning voter information into registration lists for reasons such as the volume of applications received close to Election Day and problems with the scanning equipment. To address these challenges, they reported that more staff were hired and staff worked overtime. Challenges like those experienced by the election officials we surveyed are not new. Over the last few years, we and others have found that challenges managing the voter registration process can result in eligible citizens not being registered to vote on Election Day. For example, in our October 2001 report, we found that about 46 percent of jurisdictions nationwide had problems with NVRA registrations during the November 2000 election. We cited local election officials who told us that problems processing late, incomplete, or illegible voter registration applications could result in individuals showing up at the polls and discovering they were never registered. Our survey of election officials representing the 14 jurisdictions showed that all but 1 of the jurisdictions reported removing names from registration lists during 2004 for various reasons, including that voters requested that their names be removed from the voter registration list; information from the U.S. Postal Service (USPS) showing that voters had moved outside the jurisdiction; felony records received from federal, state, or local governments identifying voters as ineligible due to felony convictions; and death records received from state or local vital statistics offices. NVRA prohibits removing a voter's name from a registration list for a federal election solely for not voting, but permits removal for certain specified reasons, including at the request of the voter, a move outside the voting jurisdiction, or due to ineligibility by reason of death or a felony conviction as provided by state law. When removing names from registration lists, election officials reported that they took various steps to ensure that the names of eligible voters were not inadvertently removed from voter registration lists. These steps included sending letters or postcards to registrants to verify that voters wanted their names removed; matching voters' identifying information with USPS data and sending voters identified by USPS as having moved outside the jurisdiction notices of removal; and matching voter registration records with felony records or death records to confirm it was the same person. Our June 2005 report showed that these same jurisdictions faced challenges verifying the accuracy of registration lists because of problems matching voter information with various records maintained by other organizations, such as records maintained by state and federal entities on felony convictions and deaths.[Footnote 11] The report echoed some of the issues highlighted in our 2001 report, wherein, on the basis of a national survey of local election officials, we reported on challenges that election officials identified with voter registration, such as obtaining accurate and timely information to update voter registration lists.[Footnote 12] Our survey of officials in the 14 election jurisdictions showed that all of the jurisdictions reported that they permitted citizens to cast provisional ballots during the November 2004 election. In addition, 12 of the 14 jurisdictions to which this was applicable reported that they offered certain first-time voters who registered by mail the opportunity to cast provisional ballots. HAVA, in general, requires that states provide an individual the opportunity to cast a provisional ballot if, for example, the individual's name does not appear on the voter registration list. HAVA also gives states discretion to determine if an individual is eligible to vote and whether the provisional ballot should be counted as a vote in accordance with state law. For example, in one state, at least a portion of a provisional ballot--e.g., votes for statewide offices--may be counted when an individual casting the ballot was eligible to vote but not at the election district where the ballot was cast. In another state, a provisional ballot is not counted if the ballot is cast in a precinct other than the individual's assigned precinct. Election officials in 13 of the 14 jurisdictions reported that 423,149 provisional ballots were cast, and 70 percent (297,662) were counted. Not all provisional votes were counted because, as election officials reported, not all provisional ballots met states' criteria for determining which ballots should be counted. Reasons that provisional ballots cast during the 2004 election were not counted, as reported by election officials, included, among others, that individuals did not meet the residency eligibility requirements, had not registered or tried to register to vote with the election office, had not submitted the voter registration applications at motor vehicle agency offices, or election officials did not have time to enter information from applicants into their voter registration lists because applications were received at the election offices very close to or after the state registration deadline. HAVA requires that states establish means of communicating how voters who cast provisional ballots can learn--free of charge--whether their votes counted. Local election officials in 12 of the 13 jurisdictions[Footnote 13] we surveyed reported that they set up mechanisms to inform voters--without cost--about the outcome of their provisional votes during the November 2004 election. These mechanisms included toll-free telephone numbers, Web sites, and letters sent to the voters who cast provisional ballots. Election officials also reported that provisional voters in their jurisdictions received written information at their polling places about how to find out the outcome of their provisional ballots, and provisional voters in 8 of the 13 jurisdictions had the opportunity to access information about the outcome of their ballots within 10 days after the election. Finally, election officials representing 8 of the 14 jurisdictions reported facing challenges implementing provisional voting for various reasons, including some poll workers not being familiar with provisional voting or, in one jurisdiction representing a large number of precincts, staff not having sufficient time to process provisional ballots. To address these challenges, the officials reported that they provided additional training to poll workers and hired additional staff to count provisional ballots. Background: The constitutional framework for elections contemplates both state and federal roles. States are responsible for the administration of both their own elections and federal elections. They also regulate various aspects of the elections process, including, for example, ballot access, registration procedures, absentee voting requirements, establishment of polling places, provision of Election Day workers, and counting and certifying the vote. Although the states are responsible for running elections, Congress has authority to affect the administration of elections. Congress' authority to regulate elections depends upon the type of election. With regard to federal elections, Congress has constitutional authority over both congressional and presidential elections. In addition, with respect to federal, state, and local elections, a number of constitutional amendments authorize Congress to enforce prohibitions against specific discriminatory acts. Federal Laws Affecting Voter Registration and Provisional Balloting: Under its various constitutional authorities, Congress has passed legislation regarding the administration of elections, including voter registration, voter registration list maintenance, and provisional voting. In 1993 Congress passed the National Voter Registration Act of 1993,[Footnote 14] also known as the "motor voter" law, to establish registration procedures designed, in part, to "increase the number of eligible citizens who register to vote in elections for Federal office. . . . protect the integrity of the electoral process. . . . [and] ensure that accurate and current voter registration lists are maintained." NVRA expanded the number of locations and opportunities for eligible citizens to apply to register to vote.[Footnote 15] Under NVRA, an eligible citizen conducting certain transactions at a motor vehicle agency--such as applying for, renewing, or changing the address on a driver's license--can simultaneously apply to register to vote in elections for federal office or, if applicable, change the voter's address. Also, various agencies, such as those involved in providing public assistance, U.S. armed forces recruitment offices, and offices in a state providing state-funded services to people with disabilities, must be designated as voter registration agencies. Each state may also designate other offices as voter registration agencies, such as public libraries, schools, offices of city and county governments, and unemployment compensation offices. In addition, NVRA requires states to accept and use a mail-in registration application developed by the Federal Election Commission. NVRA also outlined various requirements for the processing of registration forms and maintenance of voter registration lists for elections for federal office. Under NVRA, agencies designated as voter registration agencies, such as motor vehicle agencies, are to transmit completed voter registration forms to the appropriate state election officials not later than 10 days after the date of acceptance or, if accepted within 5 days before the end of the registration deadline for an election, the forms are to be transmitted not later than 5 days after acceptance. Also, under NVRA, the names of people who are registered to vote may not be removed from voter registration lists for federal elections except for the following reasons: upon written confirmation of a change of address outside the election jurisdiction; on the grounds that the voter has changed address to a location outside the election jurisdiction on the basis of change of address information from USPS, but only if there is also a failure to respond to confirmation mailings and a failure to vote in any election within two subsequent general federal elections; the request of the voter; the death of the voter; a criminal conviction, as provided by state law; and mental incapacity, as provided by state law. HAVA was enacted in October 2002 and, in addition to establishing requirements for provisional ballots for federal elections effective 2004, the law requires, in general, that individuals who registered by mail on or after January 1, 2003, and have not voted in a federal election beginning in 2004 in the state since registering to vote to show photo identification or another qualifying identification document when voting in person or by mail in a federal election. Under HAVA, the other qualifying documents that can be provided to show proof of identity and residency are a copy of a current utility bill, bank statement, government check, pay check, or other government document that shows the voter's name and address. Voter Registration in States and Local Jurisdictions: Voter registration is a prerequisite in nearly all voting jurisdictions and is based on state designated voter eligibility requirements. Ensuring that eligible persons are registered to vote is an ongoing challenge for election officials and is complicated by factors such as jurisdiction size, mobility of voters, and community diversity. Communities with large student or military populations must manage voters constantly moving in or out of a jurisdiction, and communities with diverse populations must handle substantial numbers of new citizens and face language challenges in communicating voter registration requirements. U.S. citizens who meet state voter eligibility requirements have numerous opportunities to register to vote and can do so by filling out and submitting applications by certain deadlines. States establish voting eligibility requirements that generally include that the voter is at least 18 years of age on the day of the election, a citizen of the United States, mentally competent, and meets state eligibility qualifications regarding felon status. In general, to register to vote, eligible citizens may apply at local election offices or at motor vehicle agencies and other agencies, such as public assistance agencies, covered by NVRA; by submitting applications to local election officials; or through registration drives sponsored by various groups, including political parties. Figure 1 illustrates some of the opportunities citizens have to apply to register to vote. Figure 1: Examples of Opportunities Citizens Have to Apply to Register to Vote: [See PDF for image] [A] Local election offices receive applications in different ways, such as electronically, through the mail, or delivered by hand. [End of figure] Information commonly requested on applications to vote can include full name, address, citizenship status, and signature; other information requested may include date of birth, social security number, gender, race, and/or place of birth. Figures 2 and 3 provide sample registration forms from two of the jurisdictions we visited. Figure 2: Sample of Michigan's Voter Registration Application: [See PDF for image] [End of figure] Figure 3: Sample of Texas' Voter Registration Application: [See PDF for image] [End of figure] Once an application is received and accepted, state and local jurisdictions compile registration applications into lists of registered voters. A citizen's access to voting is based primarily on the appearance of his or her name on such a list. For votes cast absentee by mail (absentee voting), for votes cast in person prior to Election Day (early voting), and for votes cast at the polls on Election Day, election officials typically verify an individual's eligibility using a list of registered voters or a poll book before allowing him or her to vote. Election officials also update and delete information from voter registration lists. However, NVRA's provisions specifically prohibit removing a name from the voter registration list for a federal election solely for failure to vote or for a change of address to another location within the same election jurisdiction. Selected Jurisdictions Reported Challenges Processing Voter Registration Applications and Various Steps They Took to Address Them: In our survey of election officials, all but one of the jurisdictions reported that they faced challenges receiving and processing voter registration applications and that they took various steps to address them. Most of these challenges occurred with applications received from voter registration drives sponsored by non-governmental organizations. Election officials in half the jurisdictions also reported challenges checking voter registration applications for completeness, or for accuracy, or for duplicates. And when asked to what extent elections staff experienced challenges entering voter information into voter registration lists, 4 of 14 jurisdictions reported this as being a challenge, to some extent. Election officials reported taking steps to address these challenges by hiring extra staff, among other things. Over the past few years, our work and the work of others have found that challenges processing voter registration applications can result in eligible citizens not being registered to vote on Election Day. Jurisdictions Reported Challenges Receiving Voter Registration Applications from Various Organizations External to Election Offices: Our survey of election officials representing the 14 jurisdictions showed that staff in most of these offices reported challenges receiving voter registration applications from various organizations external to election offices and most of these challenges occurred with applications received from voter registration drives sponsored by non- governmental organizations. Citizens in the states and jurisdictions covered by our survey can register to vote with various organizations external to election offices. In all of the jurisdictions--except those in Wisconsin, which is exempt from NVRA because it permits voters to register at the polling place on Election Day--citizens can apply to register to vote at MVAs and other NVRA agencies. Also, as mentioned earlier, citizens can register to vote during registration drives sponsored by political parties. Applications from these organizations are then sent electronically, or by mail, or are hand-delivered to election offices. Figure 4 shows the extent of challenges encountered in the 14 jurisdictions as reported by the election officials representing those jurisdictions when receiving voter registration applications from voter registration drives sponsored by non- governmental organizations, MVAs, and other NVRA agencies. Figure 4: Extent of Challenges Reported by 14 Election Jurisdictions Receiving Applications from Organizations External to Election Offices: MVAs, Other NVRA Agencies, and Voter Registration Drives Sponsored by Non-Governmental Organizations: [See PDF for image] [A] While Wisconsin is not subject to NVRA, one local jurisdiction responding to our survey--the City of Madison--received applications from other municipal agencies, such as public libraries or fire stations. [End of figure] Election officials representing the 13 jurisdictions that reported experiencing challenges to some extent or to a great extent receiving applications responded that these challenges occurred for various reasons, such as incomplete or inaccurate information on voter registration applications or applications that were received close to or after the registration deadline. In the states and 14 jurisdictions we selected for our review, deadlines for submitting applications to vote varied by state, as specified by state statute. In five of the seven states--Arizona, Michigan, New York, Texas, and Virginia-- registration applications were to be received at the election office about 1 month before the election. In one state, California, citizens could register 15 days before the election. In Wisconsin, which allows "same-day registration," eligible citizens with required identification can register to vote at the polling place on the day of the election. Election officials also reported that they took a variety of steps to try to address the challenges they encountered, including contacting individual voters to resolve issues with incomplete or inaccurate information, adding staff to handle the volume of applications received near registration deadlines, and offering training to organizations sponsoring voter registration drives. (App. III discusses in greater detail responses to our survey with regard to receiving applications from external organizations.) Election Officials in Half of the 14 Jurisdictions Reported Challenges Checking Voter Registration Applications for Completeness, Accuracy, or Duplicates: Election officials representing 7 of the 14 jurisdictions also reported that their staff experienced challenges, to some extent or a great extent, checking voter registration applications for completeness, accuracy, or duplicates. During interviews we conducted prior to the November 2004 election, state and local election officials told us that they take specific steps to check voter registration applications before entering voter information into voter registration lists. All of the officials told us that they check applications for completeness before entering voter information into voter registration lists, and with regard to accuracy, took steps to verify eligibility when an applicant applies to register. For example, officials in Arizona, Texas, Virginia, and New York said that their voter registration computer systems are programmed to calculate the age of the applicant, on the basis of the date of birth the registrant provides, and reject applications of individuals who will be younger than 18 years of age on the day of the next election. At the same time, Arizona and Michigan match their voter registration applications against state motor vehicle agency records to verify information. Also, officials we spoke with in the seven states said that they initially verify that an applicant resides in the jurisdiction. Finally, to identify duplicate registrations, election officials said that they check existing voter lists before adding new voters to the registration lists. Our survey asked election officials in the 14 jurisdictions the extent to which their offices experienced challenges checking voter registration applications for completeness, accuracy, or duplicates and officials representing 7 of the 14 jurisdictions responded that their staff experienced, to some or a great extent, challenges in at least one of these three areas. Figure 5 shows how election officials in the 14 jurisdictions responded to our survey regarding checking applications for completeness, or accuracy, or duplicates. Figure 5: Extent of Challenges Election Officials Reported Experiencing When Checking Applications for Completeness, Accuracy, or Duplicates: [See PDF for image] [End of figure] In those instances where officials responded that checking applications for completeness, accuracy, and duplication was a challenge to some or a great extent, they also identified what caused the challenges to occur and what actions they took to try to address them. According to officials in 7 of the 14 jurisdictions, challenges occurred for various reasons--including difficulties determining what information on the application was inaccurate, problems contacting individuals to obtain complete and accurate information, insufficient time between receiving the applications and the state's registration deadline, and lack of sufficient resources or staff to check the applications. Officials from these jurisdictions reported that they tried to address these challenges by, among other things, sending letters or calling applicants to get the correct information, asking individuals to provide information at the polls, and hiring additional staff or having staff work additional hours. (App. IV discusses responses to our survey regarding checking voter registration applications in greater detail.) Few of the 14 Jurisdictions Reported That They Experienced Challenges Entering Voter Information into Voter Registration Lists: Our survey also showed that few of the jurisdictions experienced challenges entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. Once election officials receive and address any issues associated with voter registration applications, they enter voter registration information into voter registration lists. In some jurisdictions with computerized registration lists, signatures of new voters are scanned into the voter registration system. When asked to what extent their staff experienced challenges entering voter information into voter registration lists, election officials representing 4 of the 14 jurisdictions responded that entering voter information into voter registration lists was, to some extent, a challenge and the remainder responded little or no challenge. When asked whether election office staff experienced challenges scanning signatures of new voters into their voter registration system, three of eight officials responded that scanning signatures was, to some extent a challenge. The remaining six officials responded that this question was not applicable to their jurisdictions. Figure 6 shows the election officials' responses to questions about entering voter information into lists, and scanning signatures of new voters into registration systems. Figure 6: Extent of Challenges Reported by Election Officials Entering Voter Information into Voter Registration Lists and Scanning Signatures of New Voters into Voter Registration Systems: [See PDF for image] [End of figure] In those instances where entering voter information into the list was, to some extent, reported as a challenge, two of the four officials responded that the challenge occurred because of the high volume of registration applications received close to Election Day--a response consistent with these officials' responses to our questions about receiving applications from organizations, like voter registration drives, external to the election office. One of these officials responded that this challenge was addressed by hiring additional staff or having staff work overtime. In those instances where scanning signatures was a challenge to some extent, one official responded that the challenge occurred because of the high volume of applications, and three officials responded that the challenge occurred because of problems with scanning equipment. Officials also responded that they tried to address this challenge by hiring additional staff or, when problems with equipment occurred, having the equipment fixed. For example, an election official in one county reported that staff faced some challenges entering voter information into the registration list and scanning new voters' signatures into the voter registration system because the election office (1) needed to hire additional staff to ensure information from voter registration applications was in the voter registration list by Election Day and (2) had problems with the scanning equipment. To address challenges associated with entering and scanning voter information, the county official reported that additional staff were hired, staff worked additional shifts, and budget requests were submitted to upgrade the scanning equipment. This official also reported that their office was challenged with having to add registration information for more than 400,000 prospective new voters to the voter registration list in the 15-day period immediately prior to the November 2004 presidential election while also having to process more than 800,000 absentee ballot applications during this same 15-day period. (App. V discusses responses to our survey regarding entering information onto voter registration lists and systems in greater detail.) Past Reports Show That Challenges Processing Voter Registration Applications Can Result in Individuals Discovering That They Were Never Registered to Vote: Our past work and the work of others has shown that challenges processing voter registration applications--including challenges receiving timely and accurate applications, checking them for completeness, and entering information into voter registration lists or systems--can have an effect on whether a citizen is registered to vote. In October 2001, we issued a report that described the operations and challenges associated with each stage of the election process, including voter registration.[Footnote 16] We reported that, on the basis of a telephone survey, an estimated 46 percent of jurisdictions nationwide had problems, in general, with NVRA registrations during the November 2000 election. Officials most frequently noted challenges with processing incomplete or illegible applications, applications that arrived late at the local election office, and applications that never arrived. We stated that, according to local officials, each of these three situations could result in individuals who show up at the polls to vote and discover that they were never registered. Other organizations have also reported that challenges processing voter registration applications can result in applicants finding that they were never registered to vote. In a 2001 report to the Congress, the Federal Election Commission (FEC) reported on the results of a survey of election officials in the 44 states and the District of Columbia subject to NVRA for the period covering 1999 through 2000 on the impact of NVRA on the administration of elections for federal office.[Footnote 17] Among other things, FEC's survey showed that some election officials expressed concerns about problems identifying multiple registrations by the same individual or applicants that provided incomplete names. Also, according to FEC, several states reported problems with election officials receiving applications from MVA offices in a timely manner. Regarding the latter, FEC stated that: "Several States reported that motor vehicle offices in some areas failed to transmit voter registration applications or changes of address to the appropriate election authorities in a timely manner. The result, unfortunately, was the effective disenfranchisement of those citizens who had duly applied but whose registrations were not processed by election day." Two years later, in another report to the Congress, FEC reported similar findings and said that some states also noted similar problems with other agencies charged with offering voter registration.[Footnote 18] FEC reiterated its conclusion about the disenfranchisement of those who had applied but whose registrations were not processed by election day and recommended, as in the earlier report, that states develop an ongoing training program for personnel in agencies covered by NVRA, such as MVAs, regarding their duties and responsibilities under NVRA. In 2001, the National Commission on Federal Election Reform, which was organized by the Miller Center of Public Affairs at the University of Virginia and The Century Foundation,[Footnote 19] expressed concerns through a task force on voter registration that delegating voter registration responsibilities to agencies and organizations outside of election offices, such as organizations that conduct voter registration drives, can result in citizens who thought they had registered being turned away from the polls on Election Day. [Footnote 20] The Commission's final report stated that voters can be turned away from the polls because of administrative errors by election offices or NVRA agencies that accept voter registration applications and noted that, although NVRA does not require it, "most states allow practically anyone to go out and register voters by taking and transmitting their mail-in applications. These people thus act in effect as deputy registrars. Election administrators we have encountered in every part of the country tell us of numerous cases where these unofficial registrars, often meaning well, mishandle or lose such applications. The applicants, of course, rightly believe they have registered. Then they show up on Election Day and find out that they are not on the list." The task force on voter registration went on to say that "third party registrars--political parties, advocacy groups and citizens groups that conduct voter registration drives--delay to send applications or fail to send them at all, sometimes inadvertently, sometimes deliberately." According to the task force, agencies other than election offices and third party registrars take applications that are incomplete or inadmissible, such as applications with postal boxes rather than street addresses, and applications that are unsigned.[Footnote 21] More recently, in May 2005, the National Task Force on Election Reform, sponsored by the non-profit Election Center,[Footnote 22] also known as the National Association of Election Officials, issued a report on various issues surrounding the 2004 general election, including voter registration. The Task Force stated that voter registration groups operated vigorously during the 2004 election and were recognized and welcomed by election officials. However, the Task Force expressed the concern that "when voter registration forms are delivered to registration authorities after the close of registration or not at all, usually at no fault of the applicant, voters find themselves not registered and not eligible to vote on Election Day."[Footnote 23] The Task Force also pointed out that NVRA agencies, which have voter registration as a secondary responsibility, are challenged to provide voter registration opportunities to their clients in a consistent manner and, although processes for transferring registrations accurately and efficiently to voter registration offices have improved, breakdowns can still occur. We did not analyze the extent to which registration applications transferred from an external organization, like a voter registration drive or MVA office, to an election office resulted in voters' names not appearing on registration lists on Election Day. However, as part of our survey of MVA officials in six states and 12 jurisdictions, we asked if any state or local election offices contacted MVA offices because individuals' names were not on the list of registered voters for the November 2004 election.[Footnote 24] Representatives of 5 of the 18 MVA offices responded that their offices were contacted by state or local election officials about individuals whose names did not appear on the registration list, but only 2 of the 5 reported data on the number of individuals affected. For example, one of representatives who reported data on the number of individuals affected said that the state election board--which has access to MVA customer data--analyzed the MVA records for 1,288 individuals whose names did not appear on voter registration rolls, but who said they had registered at MVA locations within the state. The representative reported that the board of elections found that 295 (23 percent) individuals had in fact applied to register to vote and their applications were sent to the state board of elections office. The representative also reported that, of the remaining 993 individuals, 591 (46 percent) had indicated that they did not wish to apply to register to vote and did not complete applications; 245 (19 percent) had no record of conducting MVA business or submitting voter registration applications at MVA offices; 128 (10 percent) conducted Internet, mail, or phone transactions with the MVA but did not request that paper voter registration applications be mailed to them; and 29 (2 percent) submitted their voter registration applications after the registration deadline. (App. VII discusses responses to our survey of MVA officials in 6 state and 12 local MVA offices in greater detail.) Jurisdictions Reported They Removed Voter Names from Registration Lists for Various Reasons, but Took Steps to Ensure That Eligible Voters' Names Were Not Inadvertently Removed: Election officials representing all but 1 of the 14 jurisdictions we surveyed reported that their offices removed voters' names from voter registration lists during 2004 for various reasons, including at the request of the voter and upon identification of the voter as ineligible due to a felony conviction. However, before removing names, election officials reported that their offices took steps to ensure that names of eligible voters were not inadvertently removed from lists. Our June 2005 report identified list maintenance challenges faced by election officials in the 14 jurisdictions because of problems using felony, death, and other information from existing data sources to verify the accuracy of voter registration data. Jurisdictions Reported They Took Various Steps to Ensure that Eligible Voters Were Not Inadvertently Removed from Voter Registration Lists: As discussed earlier, NVRA prohibits removing a name from a registration list for a federal election solely for non-voting but permits removal for certain specified reasons, including at the request of the voter, by reason of criminal conviction or mental incapacity under state law, the death of the voter, or upon written confirmation by the voter regarding a move out of the jurisdiction. Our survey of officials in the 14 jurisdictions showed that all but 1 of the jurisdictions said they removed names from voter registration lists during 2004 for a variety of reasons, including when the voter requested that his or her name be removed, the jurisdiction received a change of address notice from USPS, or data from other federal, state, or local organizations showed that the voter was ineligible due to a felony conviction or had died. In addition, before removing names, jurisdictions reported they took various steps to ensure that the names of eligible voters were not inadvertently removed from voter registration lists, including sending letters or postcards to voters to confirm that the voters wanted their names removed; sending follow-up notices to voters or phoning voters to notify them of removal; or matching voters' identifying information contained in voter registration records with felony records or death records maintained by other government organizations. When asked if their offices removed names of any voters from voter registration lists during 2004, 13 of 14 officials responded that, in their jurisdictions, names were removed for a variety of reasons, including when the voter requested that the name be removed from the list, felony records received from federal/state/local courts or correction facilities identified the voter as ineligible due to a felony conviction, information received from state/county vital statistics offices identified the voter as deceased, and change of address information received from USPS showed that the voter had moved outside of the jurisdiction where registered. Table 1 shows the election officials' responses to questions about reasons for removing names of voters from voter registration lists. Table 1: Election Officials' Responses to Questions about Reasons Why Voters' Names Were Removed from Voter Registration Lists in 14 Jurisdictions during 2004: [See PDF for image] [A] Delta Township, MI did not respond to this question because, as the election official reported, no names were removed from the voter registration list during 2004. [End of table] Election officials also reported that, before removing names from voter registration lists during 2004, their offices took a variety of steps to ensure that names of eligible voters were not inadvertently removed from the voter registration list. For example, election officials representing 12 of the 13 jurisdictions reported that their offices removed voters' names from voter registration lists at the request of the voter. However, before doing so, officials representing 4 of the 12 jurisdictions reported that their offices sent a letter or postcard to confirm that the voter wanted his or her name removed from the voter registration list. Representatives of 4 of the 12 jurisdictions reported that their offices took other steps including matching the voter's identifying information--e.g., name, address, date of birth, and driver's license number--with voter records to confirm it was the same person and requiring voters to file a document to indicate that they were either moving out of state or, if moving to another jurisdiction in the state, a request to transfer to another jurisdiction so that a letter can be sent to confirm the new address. Election officials representing 3 of the 12 jurisdictions reported that their offices did not take any additional steps before removing the names of eligible voters from registration lists. Election officials from 11 of the 13 jurisdictions also reported that their offices removed voters' names from registration lists because felony records received from federal, state, or local courts or correction facilities identified them as ineligible due to felony convictions. Of the 11 jurisdictions, officials representing 10 reported that, before removing names because of felony convictions, their offices matched the voter identifying information (e.g., name, date of birth, social security number, and address) contained in voter registration records with felony records to confirm it was the same person. Eight of these 11 jurisdictions also sent letters to the voters notifying them of removal on the basis of felony convictions, and 1 of the 11 also attempted to reach the voters by phone to notify them of removal. One of the 11 election officials reported that the jurisdiction did not take any additional steps before removing the names of these voters from registration lists. Regarding deceased voters, election officials in 11 of the 13 jurisdictions reported that they removed names from voter registration lists because information received from state/county vital statistics offices showed the voters as deceased. Before removing these names, all 11 of the election officials reported that they matched the voters' identifying information contained in voter registration records with death records to confirm it was the same person. According to an election official with one jurisdiction, the jurisdiction also sent a letter to the deceased voter's next of kin notifying them of removal and another election official reported that the office also asked for confirmation, and one other jurisdiction attempted to reach the voter's next of kin by phone. Election officials representing 8 of the 13 jurisdictions used obituaries as a source of information for removing voters' names from lists and 3 of the 8 reported that, before removing names of voters, they also matched voter records with state or county vital statistics information. Election officials representing 2 of the 13 jurisdictions reported that their offices relied exclusively on obituaries as a reason for removing names from voter registration lists. In addition, election officials in 10 of 13 jurisdictions reported that their offices removed names from registration lists because information provided by USPS showed that the voters had moved outside the jurisdiction where they were registered. Of the 10 jurisdictions, 9 election officials reported that their offices matched the voter's identifying information, such as name and address, contained in voter registration records with the data provided by USPS to confirm it was the same person; 8 reported that they sent a notice to the voter of the removal; and 1 reported that, in addition to matching, the office sent notices and attempted to reach voters by phone. One election official responded that the jurisdiction sends a letter to voters and voters' names are removed only after they sign and return the letter, indicating they no longer wish to be registered, and another official representing one other jurisdiction reported that, after receiving notices from voters, the office sends the voters letters asking them to confirm their current residence. Regarding the latter, the official reported that, if the address is outside of the state, voters' names are removed from the list; if it is in the same state, the information is forwarded to the receiving jurisdiction. (App. VI discusses responses to our survey on removing names from voter registration list in greater detail.) Jurisdictions Face Challenges Maintaining the Accuracy of Voter Registration Lists: Although officials reported that they took steps to ensure that names of eligible voters are not removed from registration lists, election officials still face challenges maintaining the accuracy of these lists. In June 2005, we issued a report on the efforts of state and local officials in the same seven states and 14 local jurisdictions to maintain accurate voter registration lists.[Footnote 25] We pointed out that officials faced challenges maintaining the accuracy of voter registration lists because of problems using existing data sources to verify the accuracy of voter registration data. For example: * Some state and local officials were concerned about the timeliness of death data--with the names of some deceased voters possibly remaining on registration lists after the election. Also, one state official and local officials in two states said that problems can arise in getting timely records on residents who pass away out of state. In one jurisdiction, an official expressed the concern that the election office could not identify which of several voters with the same name had died because of the lack of a birth date on some state death records. * State and local officials stated that criminal information was often incomplete, not timely, or difficult to decipher. Information on felony convictions, particularly from federal sources, was not useful because the information was old, had limited matching criteria, or was in different formats. * State and local officials were challenged ensuring voters resided within an election jurisdiction because of missing information, no street numbers in rural areas, new streets, redistricting or untimely forwarding of new addresses. We stated that some of these challenges might be addressed as state and local jurisdictions implement specific HAVA provisions that require states to (1) develop interactive computerized statewide voter registration lists for federal elections and (2) perform regular list maintenance by comparing the voter registration list against state records on felons and death.[Footnote 26] However, we also pointed out that other challenges, such as identifying duplicate registrations or deceased individuals out of state, may continue to be issues. Our June 2005 report echoed some of the issues highlighted in our 2001 report, wherein, on the basis of a national survey of local election officials, we reported on challenges that election officials identified with voter registration, such as obtaining accurate and timely information to update voter registration lists.[Footnote 27] All Jurisdictions Reported Enabling Provisional Votes to Be Cast but Not All Provisional Ballots Met States' Criteria for Whether These Votes Should Count: HAVA, in general, requires states to implement provisional voting for elections for federal office whereby individuals not listed as registered or whose eligibility is questioned by an election official must be notified about and permitted to cast a provisional ballot. Election officials representing all 14 jurisdictions reported offering provisional voting to voters, and 9 of the 14 jurisdictions informed voters that provisional ballots were available by, for example, posting signs at polling places and generating media announcements. In addition, election officials in 12 of the 14 jurisdictions reported that their jurisdictions offered certain first-time voters--those who had registered by mail after January 1, 2003, but did not provide a copy of required identification when they applied to register to vote[Footnote 28]--the opportunity to cast provisional ballots even if they were unable to provide identification at the polling place.[Footnote 29] Election officials we surveyed reported to us that, during the November 2004 election, provisional ballots were cast in 13 of the 14 jurisdictions and in 11 of the 13, the percentage of ballots counted ranged from none to nearly 90 percent. Overall, on the basis of the data reported by election officials, the 11 jurisdictions counted 70 percent of all of the provisional ballots cast. Table 2 shows the number of provisional ballots cast and counted, by jurisdiction, during the November 2004 election, as reported by election officials. Table 2: Number and Percentage of Provisional Ballots Election Officials Reported Cast and Counted in the November 2004 Election in 14 Jurisdictions: Jurisdiction: Gila County, AZ; Number of provisional ballots: Cast: 575; Number of provisional ballots: Counted: 505; Percent of provisional ballots counted: 88%. Jurisdiction: Maricopa County, AZ; Number of provisional ballots: Cast: 68,642; Number of provisional ballots: Counted: 50,719; Percent of provisional ballots counted: 74%. Jurisdiction: Los Angeles County, CA; Number of provisional ballots: Cast: 204,578; Number of provisional ballots: Counted: 166,894; Percent of provisional ballots counted: 82%. Jurisdiction: Yolo County, CA; Number of provisional ballots: Cast: 2,050; Number of provisional ballots: Counted: [A]; Percent of provisional ballots counted: [A]. Jurisdiction: City of Detroit, MI; Number of provisional ballots: Cast: 1,350; Number of provisional ballots: Counted: 123; Percent of provisional ballots counted: 9%. Jurisdiction: Delta Township, MI; Number of provisional ballots: Cast: 4; Number of provisional ballots: Counted: 1; Percent of provisional ballots counted: 25%. Jurisdiction: New York City, NY; Number of provisional ballots: Cast: 140,779; Number of provisional ballots: Counted: 77,804; Percent of provisional ballots counted: 55%. Jurisdiction: Rensselaer County, NY; Number of provisional ballots: Cast: 1,914; Number of provisional ballots: Counted: 972; Percent of provisional ballots counted: 51%. Jurisdiction: Bexar County, TX; Number of provisional ballots: Cast: 2,996; Number of provisional ballots: Counted: 602; Percent of provisional ballots counted: 20%. Jurisdiction: Webb County, TX; Number of provisional ballots: Cast: 60; Number of provisional ballots: Counted: 23; Percent of provisional ballots counted: 38%. Jurisdiction: Albemarle County, VA; Number of provisional ballots: Cast: 49; Number of provisional ballots: Counted: [A]; Percent of provisional ballots counted: [A]. Jurisdiction: Arlington County, VA; Number of provisional ballots: Cast: 147; Number of provisional ballots: Counted: 19; Percent of provisional ballots counted: 13%. Jurisdiction: City of Franklin, WI; Number of provisional ballots: Cast: 0; Number of provisional ballots: Counted: 0; Percent of provisional ballots counted: [B]. Jurisdiction: City of Madison, WI; Number of provisional ballots: Cast: 5; Number of provisional ballots: Counted: 0; Percent of provisional ballots counted: 0%. Total; Number of provisional ballots: Cast: 423,149; Number of provisional ballots: Counted: 297,662; Percent of provisional ballots counted: 70%[C]. Source: GAO survey of election officials representing 14 selected jurisdictions. [A] Respondents said they did not know how many provisional ballots were counted. Thus, the percentage of provisional votes counted could not be computed. [B] Not applicable since no provisional ballots were cast. [C] Excludes Yolo County, Calif., Albemarle County, Va., and City of Franklin, Wis. because respondents from Yolo County, Calif., and Albemarle County, Va., said they did not know how many provisional ballots were counted and no provisional ballots were cast in the City of Franklin, Wis. [End of table] To better understand what proportion of total ballots cast constituted provisional ballots cast, we compared the number of provisional ballots cast, as reported in our survey, with Web-based state and local data on total ballots cast for president for the 2004 election.[Footnote 30] Our comparison showed that the percent of provisional ballots cast, relative to total ballots cast in the 13 jurisdictions, ranged from less than 1 percent in 6 of the 13 jurisdictions to as much as 7 percent in one of the 13 jurisdictions. Table 3 shows the number of provisional ballots cast as a percent of total ballots cast, by jurisdiction, for the November 2004 election. Table 3: Number of Provisional Ballots Cast as a Percentage of Total Ballots Cast for President in the November 2004 Election in 14 Jurisdictions: Jurisdiction: Gila County, AZ; Total ballots cast[A]: 21,158; Provisional ballots cast: 575; Provisional ballots cast as a percentage of total ballots cast: 3%. Jurisdiction: Maricopa County, AZ; Total ballots cast[A]: 1,211,963; Provisional ballots cast: 68,642; Provisional ballots cast as a percentage of total ballots cast: 6%. Jurisdiction: Los Angeles County, CA; Total ballots cast[A]: 3,023,280; Provisional ballots cast: 204,578; Provisional ballots cast as a percentage of total ballots cast: 7%. Jurisdiction: Yolo County, CA; Total ballots cast[A]: 72,269; Provisional ballots cast: 2,050; Provisional ballots cast as a percentage of total ballots cast: 3%. Jurisdiction: City of Detroit, MI; Total ballots cast[A]: 326,098; Provisional ballots cast: 1,350; Provisional ballots cast as a percentage of total ballots cast: <1%. Jurisdiction: Delta Township, MI; Total ballots cast[A]: 18,060; Provisional ballots cast: 4; Provisional ballots cast as a percentage of total ballots cast: <1%. Jurisdiction: New York City, NY; Total ballots cast[A]: 2,459,653; Provisional ballots cast: 140,779; Provisional ballots cast as a percentage of total ballots cast: 6%. Jurisdiction: Rensselaer County, NY; Total ballots cast[A]: 72,983; Provisional ballots cast: 1,914; Provisional ballots cast as a percentage of total ballots cast: 3%. Jurisdiction: Bexar County, TX; Total ballots cast[A]: 475,314; Provisional ballots cast: 2,996; Provisional ballots cast as a percentage of total ballots cast: 1%. Jurisdiction: Webb County, TX; Total ballots cast[A]: 42,030; Provisional ballots cast: 60; Provisional ballots cast as a percentage of total ballots cast: <1%. Jurisdiction: Albemarle County, VA; Total ballots cast[A]: 43,726; Provisional ballots cast: 49; Provisional ballots cast as a percentage of total ballots cast: <1%. Jurisdiction: Arlington County, VA; Total ballots cast[A]: 94,650; Provisional ballots cast: 147; Provisional ballots cast as a percentage of total ballots cast: <1%. Jurisdiction: City of Franklin, WI; Total ballots cast[A]: 18,418; Provisional ballots cast: 0; Provisional ballots cast as a percentage of total ballots cast: 0%. Jurisdiction: City of Madison, WI; Total ballots cast[A]: 138,078; Provisional ballots cast: 5; Provisional ballots cast as a percentage of total ballots cast: <1%. Source: GAO survey of local election officials and Web sites of respective Secretaries of State, counties, or cities. [A] Total ballots cast does not necessarily represent the total number of votes cast and officially counted in any jurisdiction. In some jurisdictions, total ballots cast may include ballots where votes cast may or may not have been counted for a variety of reasons, including improperly marked ballots, ballots submitted without a mark for any particular candidate, ballots that included a mark for two candidates in the same election, or provisional ballots cast that were not counted. For purposes of this analysis, total ballots cast can include votes cast but not counted. For example, in Maricopa County, Arizona, the total number of ballots cast for president during the 2004 election was 1,211,963--which included 19,212 votes cast but not counted for a variety reasons--and the total number of votes cast and counted was 1,192,751. [End of table] Under HAVA, states have discretion to determine if an individual is eligible to vote and to determine whether the provisional ballot should be counted as a vote in accordance with state law. When asked to report the most frequent reasons why provisional ballots were not counted after the November 2004 election, election officials reported that the most frequent reasons were that individuals did not meet the residency eligibility requirements for the precincts where the provisional votes were cast, election officials did not have evidence that the individuals had registered or tried to register to vote before the election, and local election officials did not have evidence that the individuals casting provisional ballots had applied to register to vote at motor vehicle agency offices. Figure 7 shows the main reasons that provisional ballots were not counted after the November 2004 election in the 13 jurisdictions we surveyed where provisional ballots were cast in November 2004. Figure 7: Reasons That Provisional Ballots Were Not Counted for the November 2004 Election in Selected Jurisdictions: [See PDF for image] [A] Jurisdictions could select more than one reason. [B] Other reasons included voter did not provide acceptable identification or proof of residence; voter did not sign voter registration form; persons casting ballots were too young; voter was a convicted felon whose voting rights had not been restored; voters did not provide complete information on the voter registration applications or on the provisional ballots. [End of figure] To enable voters who cast provisional ballots in the November 2004 election to learn whether their votes counted, HAVA requires that state or local election officials establish free means of receiving the information. Local election officials in 12 of the 13 jurisdictions where provisional ballots were cast for the November 2004 election said that they set up free-access mechanisms to inform provisional voters about the outcome of their provisional votes during the November 2004 election, including toll-free telephone numbers, Web sites, and letters sent to the provisional voters who cast provisional ballots.[Footnote 31] They also said that provisional voters were provided written information at their polling places about how to find out about the outcome of their provisional ballots. Eight of the 13 jurisdictions said that provisional voters had access to information about the outcome of their ballots within 10 days after the election. Local election officials in 8 of the 14 jurisdictions said that they experienced challenges implementing provisional voting for the November 2004 election. In these jurisdictions, the election officials said that they viewed provisional voting as a challenge for various reasons, including some poll workers were not familiar with provisional voting or, in one jurisdiction representing a large number of precincts, staff reported not having sufficient time to process provisional ballots. To address these challenges, the officials reported that they provided additional training to poll workers and hired additional staff to count provisional ballots within the time frames allowed. In addition, local election officials in 3 of the 14 jurisdictions said that they experienced challenges verifying the identification of first- time voters who applied to register to vote after January 1, 2003, and who did not provide identification with their applications. Specifically, these officials said that this requirement was very time- consuming due to the high number of voters, they needed additional staff to send notices to these voters to notify them that they needed to provide the necessary identification, and the state did not have a statewide voter registration database to verify the voters' identification. (App. VII discusses responses to our survey on HAVA provisional voting and identification requirements in greater detail.) As agreed with your office, unless you publicly announce the contents of this report earlier, we plan no further distribution until 30 days after the report date. At that time we will make copies available to others on request. In addition, the report will be available at no charge on GAO's Web site at http://www.gao.gov. If you or your staffs have any questions regarding this report, please contact me at (202) 512-8777 or jenkinswo@gao.gov. Contact points for our Offices of Congressional Relations and Public Affairs may be found on the last page of this report. Key contributors are listed in appendix X. Signed by: William O. Jenkins, Jr., Director: Homeland Security and Justice Issues: [End of section] Appendix I: Scope and Methodology: Our objectives were to describe election officials' characterization of their experiences in 14 jurisdictions within 7 states with regard to the November 2004 election: (1) managing the voter registration process and any challenges related to receiving voter registration applications, checking them for completeness, accuracy, and duplication; and entering information into voter registration lists; (2) removing voter names from voter registration lists and ensuring that the names of eligible voters were not inadvertently removed; and (3) implementing provisional voting and identification requirements in accordance with HAVA and addressing any challenges encountered. In addition, we describe the experiences of selected motor vehicle agency (MVA) offices during 2004 in carrying out voter registration application procedures--specifically, assisting MVA clients with completing voter registration applications, accepting applications from these clients, and forwarding application to state or local elections offices. We first selected states and, subsequently, local jurisdictions within those states consistent with our earlier report on maintaining accurate voter registration lists.[Footnote 32] We selected the following states using a non-probability sample: Arizona, California, Michigan, New York, Texas, Virginia, and Wisconsin. Our selection of the seven states took into consideration several voter registration-related factors and unique characteristics of the states that might affect the implementation of HAVA. For example, Wisconsin has same day registration, which exempts the state from the HAVA provisional voting requirement; Arizona has on-line voter registration; and Michigan has a reputation as a model for registration practices. Also, we selected New York State because, when receiving voter registration applications or checking applications with voter registration lists, New York State may have to rely on social security number verification procedures more than other states because it has a large population who live in New York City and may not have driver's licenses. We also selected states to provide geographic diversity and variation in election administration--some administer elections at the county level and others at lower levels such as city or townships. Our goal was not to target a particular state, but rather to identify a range of issues facing states in managing the voter registration process, including updating voter registration lists, and implementing HAVA provisional voting and identification requirements. Information from these seven states cannot be generalized to all states. Table 4 describes the voter registration-related characteristics of each state. Table 4: State Selection Factors: State: Arizona; Voter registration-related characteristics: * No statewide database prior to HAVA; * Did not request a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 12.8; * Voter registration administered at county level; * Has implemented an on- line voter registration process. State: California; Voter registration-related characteristics: * Statewide database prior to HAVA that is compiled from local election lists. Local jurisdictions can access entire list; * Requested a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 26.2; * Voter registration administered at county level. State: Michigan; Voter registration-related characteristics: * Unified statewide database; * Requested a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 5.3; * Voter registration administered at township, city, and village level; * Secretary of State responsible for election and motor vehicle licensing functions. State: New York; Voter registration-related characteristics: * No statewide database prior to HAVA; * Requested a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 20.4; * Voter registration administered at county level; * Expected higher use of social security records for registrant verification due to fewer drivers in New York City. State: Texas; Voter registration-related characteristics: * Statewide database prior to HAVA that is compiled from local election lists. Local jurisdictions do not have access to entire list; * Requested a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 13.9; * Voter registration administered at county level. State: Virginia; Voter registration-related characteristics: * Unified statewide database; * Requested a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 8.1; * Voter registration administered at county level; * May use 9-digit social security number for voter registration verification (rather than 4 digits outlined by HAVA). State: Wisconsin; Voter registration-related characteristics: * No statewide database prior to HAVA. Some jurisdictions currently do not maintain voter registration rolls; * Requested a waiver from HAVA database and verification requirements deadline of January 1, 2004; * Census 2000 percent of population foreign born: 3.6; * Voter registration administered at municipal level; * Allows Election Day voter registration. Source: GAO. [End of table] Within each state, we selected the two jurisdictions using a non- probability sample. Our selection criteria for the 14 jurisdictions (2 jurisdictions in each of the 7 states) included population size, the proximity of the locations to our site visits with state election and motor vehicle officials, suggestions by state election and motor vehicle officials, and proximity to an MVA office. The local jurisdictions we selected were: * Maricopa County, Arizona; * Gila County, Arizona; * Los Angeles County, California; * Yolo County, California; * City of Detroit, Michigan; * Delta Township, Michigan; * New York City, New York; * Rensselaer County, New York; * Bexar County, Texas; * Webb County, Texas; * Arlington County, Virginia; * Albemarle County, Virginia; * City of Franklin City, Wisconsin; and: * City of Madison, Wisconsin. To address our objectives, we used a multi-pronged approach. First, we reviewed HAVA and other relevant laws, state reports, and documents related to the voter registration process in the seven states. Second, we interviewed state and local election officials in the 7 states and 14 jurisdictions to obtain information on their registration processes and implementation of the HAVA requirements for provisional voting and voter identification. Third, because so many voter registrations originate with applications from motor vehicle agencies, we also interviewed officials from these agencies in 10 jurisdictions to discuss their procedures for processing voter registration applications.[Footnote 33] In Michigan, where we spoke only with Secretary of State officials, the Secretary of State's office is responsible for elections and motor vehicle functions, such as issuing driver's licenses. Following our site visits, we sent a questionnaire to the same officials in the 14 local elections offices in the seven states to gather information about their experiences with the November 2004 election. We did not survey state election officials because voter registration and operations are carried out at the local level (e.g., county, city, or township). Consistent with our objectives, our survey focused primarily on (1) challenges, if any, processing voter registration applications, specifically, challenges receiving voter registration applications from MVAs, other NVRA agencies, and voter registration drives sponsored by non-governmental organizations; checking voter registration applications for completeness, accuracy, and duplicates; and entering voter information into registration lists or systems; (2) whether voters names were removed from voter registration lists, reasons why names were removed, and steps taken to ensure that voters names were not inadvertently removed; and (3) HAVA's provisional voting and first-time voter identification requirements, specifically, the number of provisional ballots cast and counted, including reasons, if any, why jurisdictions did not count provisional ballots; how voters were informed about provisional voting and its outcome; and whether first-time voters who registered by mail were allowed to cast provisional ballots if they did not provide identification. We also sent a survey to the state and local motor vehicle agency officials in six of the seven states[Footnote 34] to gather information about (1) their experiences assisting citizens in filling out voter registration applications and forwarding applications to election offices and (2) their experiences with the November 2004 election related to voter registration applications. Because our surveys were not sample surveys, there are no sampling errors. However, the practical difficulties of conducting any survey may introduce errors, commonly referred to as nonsampling errors. For example, difficulties in how a particular question is interpreted, in the sources of information that are available to respondents, or in how the responses we received are entered into a database or were analyzed, can introduce unwanted variability into the survey results. We took steps in the development of the surveys, the data collection, and the data analysis to minimize these nonsampling errors and help ensure the accuracy of the answers that were obtained. For example, a social science survey specialist designed the surveys in collaboration with our staff with subject matter expertise. When the data were analyzed, a second, independent analyst checked all computer programs. With regard to numbers received as part of our surveys, such as the number of provisional ballots cast and counted, we determined that the data reported via the surveys were sufficiently reliable for the purposes of this report by asking the survey respondents to identify any limitations associated with using the numbers provided and actions taken to check the accuracy of the numbers. We did not independently verify the accuracy or completeness of responses provided from our surveys of election officials and MVA officials. Local election officials and state and local MVA officials we surveyed were provided the opportunity to verify the accuracy of their responses for this report and, on the basis of the comments we received, we made technical changes, where appropriate. The results of our work cannot be generalized to state and local election jurisdictions or MVA offices either nationwide or at the state level. Appendix II contains the survey we sent to local election officials, and appendix IX contains the survey we sent to state and local officials representing MVAs. We also reviewed state and local Internet sites, such as sites maintained by Secretaries of State and County and City Clerks Offices to obtain data on total ballots cast for president for the 2004 elections in the seven states and 14 jurisdictions covered by our review. We used these data to show the relative proportion of provisional ballots cast to total ballots cast for each of the jurisdictions. These data were used for limited comparison purposes; we asked election officials to verify the accuracy of these data but we did not independently asses their reliability. In addition, we reviewed relevant federal and state statutes pertinent to the state and local jurisdictions covered by our review. To obtain information on the challenges facing state and local election officials during the registration process, we reviewed our past reports on various aspects of the election process as well as reports by the Federal Election Commission, the National Commission on Federal Reform, and a Task Force Report to Accompany the Commission's Report--both organized by the Miller Center of Public Affairs at the University of Virginia and the Century Foundation--and a May 2005 report by the National Task Force on Election Reform, sponsored by the Election Center, which is also called the National Association of Election Officials. We also reviewed various reports developed by the Election Reform Information Project, a clearinghouse for election reform information and an online forum for learning about, discussing, and analyzing election reform issues and providing research on questions of interest to the election reform community. After the November 2000 election, the Election Reform Information Project received a 3-year grant from the Pew Charitable Trusts via the University of Richmond. Our work was performed from January 2004 through September 2005 in accordance with generally accepted government auditing standards. [End of section] Appendix II: Survey of Local Election Officials: [See PDF for images] [End of section] Appendix III: Reported Experiences of Selected Local Election Offices Receiving Voter Registration Applications: This appendix summarizes the responses of local election office representatives to questions in our survey related to challenges they encountered during the November 2004 election in receiving voter registration applications from organizations external to the election office--motor vehicle agencies (MVA), other National Voter Registration Act (NVRA) agencies, and non-governmental organizations that sponsor voter registration drives; why these challenges occurred; and steps that election officials took to try to address these challenges. The statements in this appendix reflect the responses to our survey by the applicable local election officials and were not independently verified by us. Gila County, Arizona: Staff in the Gila County election office experienced little or no challenges receiving registration applications from MVA offices. However, staff faced some challenges receiving applications from NVRA agencies other than MVA offices and voter registration drives held by non-governmental organizations. These challenges occurred because applications from these sources were received after the voter registration deadline. A representative from the county election office reported that political candidates and parties that sponsored voter registration drives did not send all voter registration applications to the election office after receiving them from people who completed them. The election office tried to address these challenges by talking to the candidates and political parties about turning in applications on a timely basis. Maricopa County, Arizona: Staff in the Maricopa County election office experienced little or no challenges receiving registration applications from NVRA agencies other than MVA offices. However, staff faced some challenges receiving applications from MVA offices and faced challenges to great extent receiving applications from voter registration drives held by non- governmental organizations. Receiving applications from MVA offices and voter registration drives was challenging because the county election office received applications from these sources too close to the registration deadline. A representative from the county election office reported encountering numerous occasions when individuals who indicated that they wanted to apply to register to vote at MVA offices were not given voter registration forms to complete, and consequently these individuals were not registered to vote. Staff in the county election office tried to address these challenges by (1) adding additional staff and resources to update the voter registration list in time for the election and (2) allowing the aforementioned MVA clients who showed copies of their applications for driver's licenses indicating that they had requested to register to vote to complete voter registration forms so that they could be added to the voter registration list. Los Angeles County, California: Staff in the Los Angeles County election office experienced little or no challenges receiving registration applications from MVA and NVRA agencies. However, staff faced some challenges receiving applications from voter registration drives held by non-governmental organizations. These challenges occurred because the county election office received these applications close to or after the voter registration deadline. To address these challenges, staff from the county election office contacted these organizations and will be developing a training program focused on legal requirements for registering voters. Yolo County, California: Staff in the Yolo County election office faced little or no challenges receiving registration applications from MVA and NVRA agencies. However, staff faced some challenges receiving applications from voter registration drives held by non-governmental organizations. These challenges occurred because the county election office received the applications after the voter registration deadline, but a representative from the county election office said that the applications were still processed. City of Detroit, Michigan: Staff in the Detroit election office faced challenges to a great extent receiving applications from MVA offices, NVRA agencies, and voter registration drives held by non-governmental organizations. These challenges occurred because the applications were received either too close to or after the voter registration deadline. The representative who responded to our survey said that the election office did not try to address this challenge. Delta Township, Michigan: Staff in the Delta Township election office faced little or no challenge receiving voter registration applications from MVA and NVRA agencies. However, staff faced challenges to a great extent receiving applications from voter registration drives held by non-governmental organizations. These challenges occurred because applications were received either too close to or after the voter registration deadline and were incomplete. To address this challenge, the office added these applications to the voter registration list and sent verification letters to applicants whose applications were incomplete. If staff in the election office was not able to verify incomplete information on applicants before the election, the affected applicants' names were coded on the voter registration lists so that poll workers could ask them to identify themselves before voting. New York City, New York: Staff in the New York City election office faced little or no challenges receiving registration applications from MVA and NVRA agencies. However, staff faced some challenges receiving applications from voter registration drives held by non-governmental organizations. These challenges occurred because the applications were received too close to the voter registration deadline. To address these challenges, the city election office requested that non-governmental organizations submit all voter registration forms as soon as possible and not wait until the last minute. A representative from the city election office said that most of these groups complied with this request, but some did not. Staff in the city election office also addressed these challenges by hiring additional data entry staff to handle the large volume of registrations and by working extended hours to process the applications received before the deadline. Rensselaer County, New York: Staff in the county election office faced little or no challenges receiving registration applications from MVA offices, other NVRA agencies, and voter registration drives held by non-governmental organizations. Bexar County, Texas: Staff in the county election office experienced little or no challenges receiving applications from MVA and other NVRA agencies. However, staff faced some challenges receiving applications from voter registration drives held by non-governmental organizations. These challenges occurred because these applications were not properly completed, had missing information, had invalid addresses, or were turned in on a Friday even though the county election office had asked the organizations sponsoring voter registration drives not to do so. According to the official, the latter resulted in staff having to work overtime and weekends to process registrations. To address these challenges, staff from the county election office discussed these issues with leaders of the non-governmental organizations, state election officials, and political parties, among others. Webb County, Texas: Staff in the county election office faced some challenges receiving voter registration applications from MVA offices and faced little or no challenges receiving applications from other NVRA agencies. In addition, staff did not know or had no basis to judge whether the office faced challenges receiving applications from voter registration drives sponsored by non-governmental organizations. The challenges in receiving applications from MVA offices occurred because some applications got lost in the mail and were not received at the county election office. To address this challenge, staff reviewed information received from the Secretary of State and MVA offices, and mailed new applications to persons who had completed their applications at MVA offices. Staff also installed a lock box at the local MVA office so that MVA staff could keep completed voter registration applications in the lock box and staff from the county election office could pick up the applications at the MVA office 3 times a week. Albemarle County, Virginia: Staff in the county election office faced little or no challenges receiving voter registration applications from MVA offices. However, staff faced some challenges receiving applications from other NVRA agencies and voter registration drives held by non-governmental organizations. An official from the county election office said that these challenges occurred because (1) some NVRA agencies may not have been in compliance with NVRA and (2) the county election office may not have received applications from individuals who completed voter registration applications at a voter registration drive. To address these challenges, staff from the county election office contacted staff in NVRA agencies to encourage compliance with NVRA and offered training to all private groups who contacted the office about voter registration drives. Arlington County, Virginia: Staff in the county election office faced some challenges receiving applications from MVA offices and faced little or no challenges receiving applications from other NVRA agencies. In addition, staff faced challenges to a great extent receiving applications from voter registration drives held by non-governmental organizations. The challenges in receiving applications from MVA offices and voter registration drives occurred because the county election office received applications too close to the voter registration deadline and the county election office did not have enough staff to handle the applications that were received. To address these challenges, staff worked extra hours and the office extended the hours of temporary staff to process applications and meet deadlines. City of Franklin, Wisconsin: In response to questions about receiving applications from MVAs and NVRA agencies other than MVAs, a representative from the city election office said that this was not applicable. However, staff faced challenges to a great extent receiving applications from voter registration drives held by non-governmental organizations. These challenges occurred because (1) the city election office received these applications too close to or after the registration deadline, (2) the city election office did not have enough resources or staff to handle the applications, and (3) the applications contained inaccurate information, which required staff in the election office to do extensive follow-up work. To address these challenges, staff in the city election office attempted to make direct contact with the voters. City of Madison, Wisconsin: Staff in the city election office faced some challenges receiving voter registration applications from NVRA agencies other than MVA offices and from voter registration drives held by non-governmental organizations. These challenges occurred due to incomplete applications. In response to a question about receiving applications from MVAs, a representative from the city election office said this was not applicable. Staff in the city election office tried to address these challenges by sending letters to the applicants, if time allowed, and returning incomplete applications to staff in the NVRA agencies and non-governmental organizations sponsoring the voter registration drives. [End of section] Appendix IV: Reported Experiences of Selected Local Election Offices Checking Voter Registration Applications: This appendix summarizes the responses of local election office representatives to questions in our survey related to challenges their offices encountered during the November 2004 election with regard to checking voter registration applications for completeness, accuracy, or duplicates, and if challenges occurred, why they occurred and any actions taken to overcome them. The statements in this appendix reflect the responses to our survey by the applicable local election officials and were not independently verified by us. Gila County, Arizona: Checking voter registration applications to ensure information on the applications was complete and accurate and to identify duplicate applications posed little or no challenges to staff in the election office. Maricopa County, Arizona: Election office staff experienced some challenges checking voter registration applications to ensure the information in the applications was complete and accurate and to identify duplicate applications. These challenges occurred because local election office encountered problems contacting the individuals to obtain complete or accurate information. Staff in the local election office tried to address these challenges by putting the applications in a suspense file and sending letters to the applicants requesting the missing information. Los Angeles County, California: While election staff experienced little or no challenges checking individual applications for completeness and duplicates, they experienced challenges, to some extent, checking the accuracy of voter registration applications because the elections office did not have sufficient time, resources, and staff to check them. To address this challenge, additional staff were hired, multiple work shifts were added, and a great deal of overtime was worked. Yolo County, California: Election office staff experienced little or no challenges checking voter registration applications for accuracy, completeness, or duplicates. City of Detroit, Michigan: Election office staff experienced challenges, to a great extent, checking voter registration applications for completeness, accuracy, and duplicates. The challenges they faced checking voter registration applications were due to problems with (1) determining what information was inaccurate, (2) contacting individuals to obtain complete or accurate information, and (3) checking the accuracy of information in the applications against various databases, such as motor vehicle or Social Security Administration databases, or against state records. To overcome this challenge, election office staff attempted to reach the applicants by telephone and mailed response cards to them to verify the information contained in their applications. Delta Township, Michigan: Although checking applications for accuracy or duplicates posed little or no challenge to election office staff, they experienced challenges, to some extent, checking the completeness of voter registration applications. These challenges occurred because hundreds of registration forms were received too close to the registration deadline and staff encountered difficulties determining what information was not accurate. To address these challenges, election staff added the names of the applicants to the voter registration lists and sent the applicants letters asking them to verify the information contained in their registration applications. If the applicants did not respond to the letters, their names were placed on the registration list and if they appeared at the polls to vote, they were asked to provide identification. New York City, New York: Election office staff faced little or no challenges checking voter registration applications to ensure information on the applications was complete and accurate and to identify duplicate applications. Rensselaer County, New York: Election office staff faced little or no challenges checking voter registration applications to ensure information on the applications was complete and accurate and to identify duplicate applications. Bexar County, Texas: Election office staff faced little or no challenges checking voter registration applications to ensure information on the applications was complete and accurate and to identify duplicate applications. Webb County, Texas: Election staff experienced little or no challenges checking voter registration applications to ensure information on the applications was complete and to identify duplicate applications. The official who responded to our survey did not know the extent of challenges, if any, that staff may have experienced checking voter registration applications for accuracy. Albemarle County, Virginia: Election office staff faced little or no challenges checking voter registration applications to ensure information on the applications was complete and accurate and to identify duplicate applications. Arlington County, Virginia: Election office staff experienced some challenges checking voter registration applications for completeness, accuracy, and duplicates. These challenges occurred because the election office did not have enough resources and staff to check the applications, and staff did not have enough time to check the applications between the registration deadline and when voting began. To address these challenges, staff worked extra hours. City of Franklin, Wisconsin: To a great extent, election office staff experienced challenges checking applications for completeness, accuracy, and duplicates. Staff experienced these challenges because of (1) difficulties determining what information was inaccurate, (2) problems contacting individuals to obtain complete or accurate information, and (3) insufficient time between the state's registration deadline and when voting began to check whether applications were complete, accurate, or duplicates. To overcome these challenges, staff from the election office attempted to make direct contact with the voters. City of Madison, Wisconsin: To some extent, election office staff faced challenges checking applications for completeness and accuracy but had little or no problems checking applications for duplicates. The challenges of checking the completeness and accuracy of applications occurred because of problems with contacting individuals to obtain complete and accurate information. To overcome these challenges, additional staff were hired. [End of section] Appendix V: Reported Experiences of Selected Local Election Offices Entering Information into Voter Registration Lists: This appendix summarizes the responses of local election office representatives to questions in our survey related to challenges they encountered during the November 2004 election entering information on eligible voters into voter registration lists, and when challenges occurred, why they occurred and actions taken to overcome them. The statements in this appendix reflect the responses to our survey by the applicable local election officials and were not independently verified by us. Gila County, Arizona: Entering information on eligible voters into voter registration lists posed little or no challenges to staff in the election office. However, scanning signatures of new voters into the voter registration system posed challenges to some extent because staff in the election office encountered problems with the scanning equipment. To address the challenges of scanning new voters' signatures into the voter registration system, the equipment was repaired. Maricopa County, Arizona: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. Los Angeles County, California: To some extent, election office staff experienced challenges entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. These challenges occurred because the election office lacked sufficient resources and staff and had problems with the scanning equipment. To address these challenges, additional staff were hired, and staff worked additional shifts and a great deal of overtime. In addition, budget requests were submitted to upgrade the scanning equipment. Yolo County, California: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. City of Detroit, Michigan: To some extent, election office staff experienced challenges scanning signatures of new voters into the voter registration system but experience little or no challenges entering information on eligible voters into the voter registration lists. Scanning signatures of new voters into the voter registration system was a challenge because there were too many applications to enter into the system by the registration deadline or Election Day and because of problems with the scanning equipment. To address this challenge, election office staff scanned as many signatures as possible and waited until the equipment malfunctioning could be corrected. Delta Township, Michigan: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. New York City, New York: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. Rensselaer County, New York: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists and scanning signatures of new voters into the voter registration system. Bexar County, Texas: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists. Scanning signatures of new voters into the voter registration system was not applicable to this election office. Webb County, Texas: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists. Scanning signatures of new voters into the voter registration system was not applicable to this election office. Albemarle County, Virginia: Election office staff experienced challenges, to little or no extent, entering voter information into voter registration lists. Scanning signatures of new voters into the voter registration system was not applicable to this election office. Arlington County, Virginia: Election office staff experienced challenges, to some extent, entering information on eligible voters into voter registration lists because the election office lacked sufficient resources and staff to enter the registration application information into these lists by registration deadline or Election Day. To address these challenges, election office staff worked extra hours to meet deadlines for entering information on eligible voters into the voter registration lists. Scanning signatures of new voters into the voter registration system was not applicable to this election office. City of Franklin, Wisconsin: To some extent, election office staff experienced challenges entering information on eligible voters into voter registration lists because of the volume of registration applications that had to be entered by the state's registration deadline or Election Day. To address these challenges, staff worked extended hours and additional staff was hired to process registration applications that were required to be entered into voter registration lists by Election Day. Scanning signatures of new voters into the voter registration system was not applicable to this election office. City of Madison, Wisconsin: To some extent, election office staff experienced challenges entering information on eligible voters into voter registration lists because of the volume of registration applications that had to be entered by the state's registration deadline or Election Day. To address these challenges, staffing was increased and the office operated with extended hours. Scanning signatures of new voters into the voter registration system was not applicable to this election office. [End of section] Appendix VI: Reported Experiences of Selected Local Election Offices Removing Names from Voter Registration Lists: This appendix describes the responses of local election officials, representing 14 jurisdictions on their reported experiences removing the names of voters from voter registration lists during 2004. All but 1 of the 14 local election officials responding to our survey reported that voters' names were removed from registration lists during 2004 for various reasons, such as voter requested name be removed or records identified voter as deceased or as ineligible due to a felony conviction. For information on the reasons that election officials in each jurisdiction removed names of voters from registration lists, see table 1. The local election officials also reported steps that they took before removing voters' names from the voter registration lists in order to ensure that names of eligible voters were not inadvertently removed. The statements in this appendix reflect the responses to our survey by the applicable local election officials and were not independently verified by us. Gila County, Arizona: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * sent the voter a follow-up notice about removal, because the voter failed to respond to a notice from the voter registrar and had not voted or had not appeared to vote in the most recent two federal elections, but received no response to the follow-up notice; * sent the voter a notice of removal because change-of-address information from the U.S. Postal Service showed that the voter had moved to another jurisdiction or felony records identified the voter as ineligible due to a felony conviction; * sent a letter or postcard to the voter to confirm that the voter wanted his/her name removed from the voter registration list as requested; * sent a letter to the voter notifying of removal because court records identified the voter as ineligible due to mental incompetence; * sent a letter to the deceased voter's next of kin notifying of removal and asking for confirmation of the voter's death; * contacted the funeral home to obtain the deceased voter's identifying information (such as name, date of birth, and address); * matched information from a newspaper obituary with information from the state/county vital statistics offices; and: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address (i.e., the U.S. Postal Service National Change of Address list), felony convictions, mental incompetence, and deaths. Maricopa County, Arizona: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * mailed two notices to the voter--one to the mailing address on file and one to the residence address, if different from the mailing address--when the office received information indicating that the voter had moved from the address on file; if both notices were returned undeliverable, moved the voter to an inactive registration list and if no voter activity occurred through two federal elections, the voter's registration was canceled; * canceled the voter's registration when the voter requested in writing that his/her name be removed from the registration list; * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address (i.e., the U.S. Postal Service National Change of Address list), felony convictions, mental incompetence, and deaths; * sent a letter to the voter asking if he/she wanted to remain registered in the county when information from the U.S. Postal Service indicated that the voter had moved to another jurisdiction, and if the voter signed and returned the letter indicating that he/she no longer wanted to be registered in the county, the registration was canceled; * sent a letter to the voter notifying of removal because court records identified the voter as ineligible due to a felony conviction and gave the voter an opportunity to inform the election office that the information provided from the courts was incorrect or that the felony was overturned; and: * sent a letter to the deceased voter's family asking to confirm the voter's death when information provided by the state/county vital statistics offices or newspaper obituaries was insufficient to positively identify the voter as deceased. Los Angeles County, California: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address (e.g., the U.S. Postal Service National Change of Address list), felony convictions, mental incompetence, and deaths; * sent the voter a letter and attempted to reach the voter by phone to notify of removal on the basis of a felony conviction; and: * sent the voter a letter notifying of removal because court records identified the voter as ineligible due to mental incompetence. Yolo County, California: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address (i.e., the U.S. Postal Service National Change of Address list), felony convictions, and deaths; * sent the voter a letter or notice of removal to confirm that the voter wanted his/her name removed from the voter registration list as requested or to notify the voter of removal on the basis of change-of- address information from the U.S. Postal Service; and: * matched information from newspaper obituaries with information from state/county vital statistics offices. City of Detroit, Michigan: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address and deaths; * sent the voter a follow-up notice about removal because the voter failed to respond to a notice from the voter registrar and had not voted or had not appeared to vote in the most recent two federal elections, but received no response to the follow-up notice; * sent the voter a notice of removal because the voter no longer resided in the jurisdiction; and: * canceled the voter's registration as requested by the voter and annotated the registration card with the reason the registration was canceled. No steps were taken before removing a voter's name from the voter registration list when felony records identified the voter as ineligible due to a felony conviction. New York City, New York: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address, felony convictions, mental incompetence, and deaths; * sent the voter a notice of removal because change-of-address information from the U.S. Postal Service showed that the voter had moved to another jurisdiction, or felony records identified the voter as ineligible due to a felony conviction, or court records identified the voter as ineligible due to mental incompetence; and: * sent the voter a letter or postcard to confirm that the voter wanted his/her name removed from the voter registration list as requested. Rensselaer County, New York: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * sent the voter a notice of removal because change-of-address information from the U.S. Postal Service showed that the voter had moved to another jurisdiction and: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address, felony convictions, or deaths. No steps were taken before removing voters' names from the voter registration list when voters requested that names be removed or newspaper obituaries identified voters as deceased. Bexar County, Texas: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * sent the voter a follow-up notice about removal because the voter failed to respond to a notice from the voter registrar and had not voted or had not appeared to vote in the most recent two federal elections, but received no response to the follow-up notice; * sent a letter or postcard to the voter to confirm that the voter wanted his/her name removed from the voter registration list as requested; * sent the voter a letter notifying of removal on the basis of a felony conviction or mental incompetence; and: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on felony convictions, mental incompetence, and deaths. Webb County, Texas: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * sent the voter a follow-up notice about removal because voter failed to respond to a notice from the voter registrar and had not voted or had not appeared to vote in the most recent two federal elections, but received no response to the follow-up notice; * sent the voter a notice of removal because change-of-address information from the U.S. Postal Service showed that the voter had moved to another jurisdiction; * sent the voter a letter notifying of removal on the basis of a felony conviction; * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on felony convictions and deaths; * attempted to reach the deceased voter's next of kin by phone to notify of removal and to ask for confirmation of voter's death; and: * matched information from a newspaper obituary with information from state/county vital statistics offices. Albemarle County, Virginia: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * required the voter to file a document with the election office before any action was taken to remove the name from registration list as the voter requested; * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on felony convictions, mental incompetence, deaths, or change of registration notice from another state; and: * sent the voter a letter notifying of removal on the basis of a felony conviction. No steps were taken before removing voters' names from the voter registration lists when newspaper obituaries identified voters as deceased. Arlington County, Virginia: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * matched the voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from various sources on changes of address, felony convictions, mental incompetence, and deaths; * sent a letter asking the voter to confirm current residence because change-of-address information from the U.S. Postal Service showed that the voter had moved to another jurisdiction; * sent the voter a postcard to confirm a request to remove his/her name from the voter registration list; and: * sent the voter a letter notifying of removal on the basis of a felony conviction or on the basis of court records that identified the voter as ineligible due to mental incompetence. City of Franklin, Wisconsin: Before removing names from the voter registration list, the following steps were taken to ensure that names of eligible voters were not inadvertently removed: * matched voter's identifying information (such as name, address, date of birth, and social security number) contained in the voter registration list with information received from the U.S. Postal Service National Change of Address list; * sent the voter a notice of removal because change-of-address information from the U.S. Postal Service showed that the voter had moved to another jurisdiction; and: * attempted to reach the voter by phone to confirm that he/she had moved outside the jurisdiction. No steps were taken before removing voters' names from the voter registration list when voters requested that names be removed or newspaper obituaries identified voters as deceased. City of Madison, Wisconsin: When the office received voter cancellation cards from other municipalities, staff checked the voters' dates of birth before removing their names from the voter registration list to ensure that names of eligible voters were not inadvertently removed from the list. No steps were taken before removing voters' names from the voter registration lists when voters requested that their names be removed or newspaper obituaries identified voters as deceased. [End of section] Appendix VII: Reported Experiences of Selected Local Election Offices Implementing Provisional Voting: This appendix summarizes the responses of local election office representatives to questions in our survey and interviews about their experiences during the November 2004 election in implementing the requirements in HAVA for provisional voting and identification requirements for certain first-time voters who applied to register to vote after January 1, 2003. Unless otherwise noted, the statements in this appendix reflect the responses to our survey by the applicable local election officials and were not independently verified by us. Gila County, Arizona: In Gila County, 575 provisional ballots were cast and 505 of these ballots--88 percent--were counted. For a provisional vote to count, the voter had to be a qualified voter in the precinct. The main reasons for not counting the other 70 provisional ballots were that (1) the voters did not meet the residency eligibility requirements, (2) the election office received voter registration applications very close to or after the registration deadline, and (3) the provisional ballots or the envelopes containing those ballots were incomplete or illegible. The county election office did not experience challenges implementing the provisional voting requirements during the November 2004 election. The county election office used various means to inform voters of the availability of provisional ballots and the outcome of provisional votes. Representatives we spoke with at the county election office told us that they notified voters of the availability of provisional ballots by posting signs at polling places. Voters were also provided with written information at their polling places informing them that they could find out whether or not their provisional votes were counted and, if the votes were not counted, the reason they were not counted. Voters who cast provisional ballots could find out the outcome of their votes by calling a toll-free telephone number, calling the local election office, calling the state election office, and via a letter from the local election office informing voters of the outcome of their provisional votes. Information on the outcome of the provisional votes was made available to the voters between 6 and 10 days after the election. In response to a question in our survey asking if the county allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place, the representative who responded to the survey said this was not applicable. Arizona passed a ballot initiative in 2004 requiring voters to show identification prior to receiving ballots.[Footnote 35] The county election office did not report experiencing challenges verifying the identification of these first- time voters. Maricopa County, Arizona: In Maricopa County, 68,642 provisional ballots were cast and 50,719 of these ballots--74 percent--were counted. For a provisional vote to count, the voter had to be a qualified voter in the precinct. The main reasons for not counting the other 17,923 provisional ballots were that (1) no evidence existed that the individuals who cast these ballots had registered or tried to register to vote directly with the elections office, (2) the voters did not meet the residency eligibility requirement, and (3) other reasons including election officials not receiving voter registration applications until after the state's deadline to register to vote. In describing challenges that the county election office faced in implementing provisional voting for the November 2004 election, officials from the county election office said they had to modify the form and envelope used for provisional voting to provide a statement for the voter to sign to indicate U.S. citizenship. Voters were informed of the availability of provisional ballots and outcome of provisional votes using various means. Representatives we spoke with at the county election office told us that they notified voters of the availability of provisional ballots by posting signs at polling places. Voters were also provided with written information at their polling places informing them that they could find out whether or not their provisional votes were counted and, if the votes were not counted, the reason they were not counted. Voters who cast provisional ballots could find out the outcome of their votes by calling a toll- free telephone number, checking a Web site, calling the local election office, calling the state election office, visiting the local election office, and via a letter from the local election office informing voters of the outcome of their provisional votes. Information on the outcome of the provisional votes was made available to the voters between 6 and 10 days after the election. In response to a question in our survey asking if the county allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place, the representative who responded to the survey said this was not applicable. Arizona passed a ballot initiative in 2004 requiring voters to show identification prior to receiving ballots. The county election office did not report experiencing challenges verifying the identification of these first-time voters. Los Angeles County, California: In Los Angeles County, 204,578 provisional ballots were cast during the November 2004 election and 166,894 of these ballots--82 percent--were counted. Provisional votes are counted when they are cast by qualified voters in any precinct within the county. The main reasons for not counting the other 37,684 ballots were that (1) no evidence existed that the voters had registered or tried to register to vote, (2) the voters did not sign the voter registration forms, and (3) the voters did not provide dates or places of birth on the voter registration forms or the provisional ballot envelopes. The county election office experienced challenges implementing the HAVA provisional voting requirements during the November 2004 election and took steps to address these challenges. Provisional voting was a challenge because staff had to prepare duplicate ballots to remove ineligible or invalid contests when voters cast their provisional ballots at the wrong precinct. To overcome this challenge, staffing was increased to prepare the duplicate ballots. Voters were informed of the availability of provisional ballots and how to find out the outcome of provisional votes using various means. Representatives we spoke with at the county election office told us that voters were notified of the availability of provisional ballots by posting signs at polling places, conducting public education campaigns, and working with community-based organizations. Also, they said that information on provisional voting is contained in California's Voting Bill of Rights and the state's election code. Voters were provided written information at their polling places informing them that they could find out the outcome of their provisional votes by calling a toll- free telephone number or the local election office. Information on the outcome of the provisional votes was made available to the voters more than 10 days after the election. The county allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place; 12,264 of these first-time voters cast provisional ballots during the November 2004 election. To alert pollworkers and elections staff of first-time voters who were required to provide identification at the polls because they did not provide it when they registered by mail, notations were made to the voter file, voter roster, and absentee identification envelope. The county election office reported experiencing challenges verifying the identification of these first-time voters. Specifically, this proved to be a challenge because the state did not have a statewide voter registration database to verify the identification of these voters. To overcome this challenge, voters were requested to provide identification at the polls and absentee voters were requested to enclose identification with their ballots. Yolo County, California: In Yolo County, 2,050 provisional ballots were cast during the November 2004 election. The representative from the election office who responded to our survey did not know how many of these provisional ballots were counted. For a provisional vote to count, the voter had to be a qualified voter in any precinct within the county. The most frequently cited reason for not counting provisional ballots were (1) a lack of evidence that the voters had submitted voter registration applications at MVA offices, (2) voters not providing identification as required by HAVA for individuals who registered by mail and were voting for the first time in the precinct or jurisdiction, and (3) voters not signing a sworn statement that they met the qualifications to be eligible to vote in the precinct or jurisdiction. The county election office did not experience challenges implementing the HAVA provisional voting requirement during the November 2004 election. The county election office used various means to inform voters of the availability of provisional ballots and outcome of provisional votes. Representatives we spoke with at the county election office told us that voters were notified of the availability of provisional ballots by posting signs at polling places and mailing sample ballots to voters. Voters were also provided written information at their polling places informing them that they could find out the outcome of their provisional votes by calling a toll-free telephone number or the local election office. Information on the outcome of the provisional votes was made available to the voters more than 10 days after the election. The county allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place. The representative from the county election office who responded to our survey did not know how many first-time voters cast provisional ballots during the November 2004 election. To alert pollworkers and elections staff about first- time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The county election office did not report experiencing challenges verifying the identification of these first-time voters. City of Detroit, Michigan: In Detroit, 1,350 provisional ballots were cast during the November 2004 election, and 123 of these ballots--9 percent--were counted. Provisional votes are counted when they are cast by qualified voters in their assigned precincts. The main reasons for not counting the other 1,227 provisional ballots were that the voters did not (1) provide acceptable identification or proof of residence, (2) sign sworn statements that they met the qualifications to be eligible to vote in the precincts or jurisdictions, or (3) meet the residency eligibility requirements for the precincts or jurisdictions. The election office experienced challenges implementing provisional voting during the November 2004 election. Provisional voting was a challenge because Detroit has the largest number of voting precincts in Michigan, and the 6-day time frame for processing provisional ballots was very challenging and unrealistic. To overcome this challenge, the entire department's employees were mobilized to process provisional ballots. The election office used various means to inform voters of the availability of provisional ballots and outcome of provisional votes. Representatives we spoke with at the election office told us that election staff notified voters of the availability of provisional ballots by speaking to community and church groups and through a voter education program. Voters were provided with written information at their polling places informing them that they could find out the outcome of their provisional votes by contacting the local election office by phone and in person, and by receiving a letter from the election office. Information on the outcome of the provisional votes was made available to the voters within 6 to 10 days after the election. The city allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place. The representatives from the city election office who responded to our survey did not know how many of these first-time voters cast provisional ballots during the November 2004 election. To alert poll workers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The city election office did not report experiencing challenges verifying the identification of these first-time voters. Delta Township, Michigan: In Delta Township, four provisional ballots were cast during the November 2004 election, and one of these ballots was counted. Provisional votes are counted when they are cast by qualified voters in their assigned precincts. The reason for not counting the other three ballots was that the voters' registration applications were postmarked after the registration deadline. The township election office experienced challenges implementing the HAVA provisional voting requirement during the November 2004 election and took steps to address these challenges. Provisional voting was a challenge because it was viewed as too complex and time-consuming for poll workers to administer along with all the other issues they face on Election Day. To overcome this challenge, the election office provided the poll workers with general information on provisional voting and instructed staff in precincts to call the election office about each provisional envelope ballot being considered. The election office used various means to inform voters of the availability of provisional ballots and outcome of provisional votes. Representatives we spoke with at the election office told us that voters in the township were notified of the availability of provisional ballots primarily through the media. Voters were also provided with written information at their polling places informing them that they could find out the outcome of their provisional votes by calling a toll- free telephone number, calling the local or state election office telephone number, or visiting the local election office in person. Information on the outcome of the provisional votes was made available to the voters within 5 days after the election. Delta Township allowed first-time voters who registered by mail to cast provisional ballots if they were not able to provide identification at the polling place. To alert pollworkers and elections staff about first- time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. None of these first-time voters cast provisional ballots during the November 2004 election. The county election office reported experiencing challenges verifying the identification of these first- time voters. Specifically, this was viewed as very time-consuming due to the high volume of voters, and that other voters suffered because this resulted in delays. To overcome these challenges, poll workers asked for identification as needed. New York City, New York: In New York City, 140,779 provisional ballots were cast and 77,804 of these ballots--55 percent--were counted. Provisional votes are counted when they are cast by qualified voters in their precincts. The main reasons for not counting the other 62,975 provisional ballots were that (1) no evidence existed that the individuals who cast these ballots had registered or tried to register to vote directly with the elections office, (2) the envelopes or ballots were incomplete or illegible, and (3) voters did not meet the residency eligibility requirements. The city election office did not report experiencing challenges with implementing provisional voting. In response to a question about how the city election office notified voters of the availability of provisional ballots, election office representatives said that voters were already familiar with the system.[Footnote 36] Voters were not provided with written information at their polling places informing them that they could find out whether or not their provisional votes were counted and, if the votes were not counted, the reason they were counted. Instead, the election office sent letters to only those provisional voters whose ballots were not counted because state election law requires notifying only those voters whose provisional ballots were not counted. Information on the outcome of the provisional ballots was made available to the voters more than 10 days after the election. The city allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place; 551 of these first-time voters cast provisional ballots during the November 2004 election. The names of these voters were marked on the registration list so that poll workers knew that these voters needed to show identification. The county election office did not report experiencing challenges verifying the identification of these first-time voters. Rensselaer County, New York: In Rensselaer County, 1,914 provisional ballots were cast and 972 of these ballots--51 percent--were counted. For a provisional vote to count, the voter had to be a qualified voter in the precinct. The main reasons for not counting the other 942 provisional ballots were that (1) the county election office received registration applications after the state's registration deadline, (2) no evidence that individuals had registered or tried to register directly with the election office, and (3) voters did not meet residency eligibility requirements. The county election office did not report experiencing challenges with implementing provisional voting. In response to a question about how the county election office notified voters of the availability of provisional ballots, election office representatives said that poll worker inspectors did so if they were not listed in the voter rolls. Voters were not provided with written information at their polling places informing them that they could find out whether or not their provisional votes were counted and, if the votes were not counted, the reason they were not counted. Voters who cast provisional ballots could find out the outcome of their votes by calling the election office. Information on the outcome of the provisional votes was made available to the voters more than 10 days after the election. The county allowed first-time voters who registered by mail as of January 2003 but did not provide a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place. None of these first-time voters cast provisional ballots during the November 2004 election. The names of these voters were marked on the registration list so that poll workers knew that these voters needed to show identification. The county election office did not report experiencing challenges verifying the identification of these first-time voters. Bexar County, Texas: In Bexar County, 2,996 provisional ballots were cast during the November 2004 election and 602 of these ballots--20 percent--were counted. Provisional votes are counted when they are cast by qualified voters in their assigned precincts. The main reasons that the other 2,394 provisional ballots were not counted were that (1) voters did not meet the residency eligibility requirements for the precinct or jurisdiction and (2) no evidence existed that the individuals had registered or tried to register directly with the elections office or had submitted their voter registration applications at MVA offices. The county election office reported experiencing minor challenges implementing provisional voting during the November 2004 election, which included missing signatures and provisional ballots cast in incorrect precincts. To address this issue, efforts were made to review all available records for the missing or incorrect information, such as the voter registration database, secretary of state records, motor vehicle agency records, and cards containing the voters' original signatures. The election office used various means to inform voters of the availability of provisional ballots and outcome of provisional votes, such as posting signs at polling places. This information was also made available to voters upon request. Voters were not provided with written information at their polling places informing them that they could find out whether or not their provisional votes were counted and, if the votes were not counted, the reasons they were not counted. Voters could find out the outcome of their provisional votes by calling the local election office or by receiving a letter from the election office. Information on the outcome of the provisional votes was made available to the voters within 6 to 10 days after the election. The county allowed first-time voters who registered by mail after January 2003 and did not send a copy of identification with their applications to cast provisional ballots if they were not able to provide identification at the polling place. The representative from the county election office did not know how many first-time voters cast provisional ballots during the November 2004 election. To alert poll workers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The county election office did not report experiencing challenges verifying the identification of these first-time voters. Webb County, Texas: In Webb County, 60 provisional ballots were cast during the November 2004 election and 23 of these ballots were counted. Provisional votes are counted when they are cast by qualified voters in their assigned precincts. The main reasons for not counting the other 37 ballots were (1) voters did not meet the residency eligibility requirements for the precinct or jurisdiction, (2) a lack of evidence that the individuals had submitted their voter registration applications at MVA offices, and (3) the county election office received voter registration applications after registration deadline. The county election office did not report experiencing challenges with implementing provisional voting. The election office provided written information to voters who cast provisional ballots informing them that they would receive a letter from the election office notifying them of the outcome of their votes. Information on the outcome of the provisional votes was made available to the voters in 5 days or less after the election. The county allowed first-time voters who registered by mail after January 2003 but did not provide a copy of identification with their registrations to cast provisional ballots if they were not able to provide identification at the polling place. None of these first-time voters cast provisional ballots during the November 2004 election. To alert pollworkers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The county election office did not report experiencing challenges verifying the identification of these first-time voters. Albemarle County, Virginia: In Albemarle County, 49 provisional ballots were cast during the November 2004 election. The representative from the county election office who responded to our survey did not know how many provisional ballots were counted. For provisional votes to be counted, they must be cast by qualified voters in their assigned precincts. The main reasons for not counting provisional votes were that (1) voters did not meet the residency eligibility requirements for the precinct or jurisdiction because they had moved from one jurisdiction to another within Virginia and had not updated their voter registration records within the required time frame and (2) no evidence existed that the individuals had submitted their voter registration applications at MVA offices or at other NVRA agencies. The county election office reported experiencing challenges implementing provisional voting. Specifically, provisional voting was viewed as a challenge because ballots were provided to ineligible voters to placate them, given the common misperception among some voters that they could cast ballots anywhere. In addition, because various groups created a high level of mistrust among voters, who at times were confrontational or disruptive, provisional ballots were provided to placate voters and to keep order at the polls. To address these challenges, the county election office said it would expand outreach efforts to ensure individuals who were moving understood the need to provide the election office with a timely change of address. The county election office used various means to inform voters of the availability of provisional ballots and outcome of provisional votes. Representatives we spoke with at the county registrar office told us that voters were notified of the availability of provisional ballots through the media. Voters were also provided with written information at their polling places informing them that they could find out the outcome of their provisional votes by calling a toll-free telephone number, calling the local or state election office, contacting the local election office in person, or receiving a letter from the local election office. Information on the outcome of the provisional votes was made available to the voters within 6 to 10 days after the election. The county allowed first-time voters who registered by mail after January 2003 but did not provide a copy of identification with their registrations to cast provisional ballots if they were not able to provide identification at the polling place. None of these first-time voters cast provisional ballots during the November 2004 election. To alert pollworkers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The county election office did not report experiencing challenges verifying the identification of these first-time voters. Arlington County, Virginia: In Arlington County, 147 provisional ballots were cast during the November 2004 election and 19 of these ballots--13 percent--were counted. Provisional votes are counted when they are cast by qualified voters in their assigned precincts. The main reasons for not counting the other 128 ballots were (1) a lack of evidence that the individuals had registered or tried to register directly with the elections office, (2) a lack of evidence that the individuals had submitted their voter registration applications at MVA offices, and (3) that they did not meet the residency eligibility requirements for the precinct or jurisdiction. The county election office reported experiencing challenges implementing provisional voting. Specifically, provisional voting was viewed as a challenge because some poll workers did not understand, according to the county election office, that anyone who insisted on voting a provisional ballot was entitled to do so and, as a result, some voters whose names were not in the registration list might not have been offered provisional ballots. Provisional voting was also a challenge because, according to the county election office, lawyers stationed outside the polls sometimes erroneously told voters that they could insist on voting a provisional ballot even though poll workers had directed them to their correct polling places. The election office plans to address some of these issues during training sessions for poll workers. The county registrar office used various means to inform voters of the availability and outcome of provisional votes. Representatives we spoke with at the county registrar office told us that voters were notified of the availability of provisional ballots by posting signs at polling places. Voters were also provided with written information at their polling places informing them that they could find out the outcome of their provisional votes by calling a toll-free telephone number, calling the local or state election office telephone number, contacting the local election office in person, or by receiving a letter from the local election office. Information on the outcome of the provisional votes was made available to the voters within 6 to 10 days after the election. The county allowed first-time voters who applied to register to vote by mail after January 2003 but did not provide a copy of identification with their registrations to cast provisional ballots if they were not able to provide identification at the polling place. None of these first-time voters cast provisional ballots during the November 2004 election. To alert pollworkers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The county election office did not report experiencing challenges verifying the identification of these first-time voters. City of Franklin, Wisconsin: No provisional ballots were cast in Franklin during the November 2004 election. If provisional ballots had been cast, voters would have been provided written information at their polling places informing them that they could find out the outcome of their provisional votes. Provisional votes would have been counted if the ballots were cast by qualified voters in their assigned precincts. The election office reported experiencing challenges implementing provisional voting and took steps to address these challenges. Specifically, providing thorough instructions to all election workers and voters about provisional voting during an extremely high-turnout election was challenging. To address this issue, the city election office provided thorough and repeated poll worker education. Provisional ballots were made available to individuals who applied to register to vote after January 2003 but did not send a copy of identification with their applications. None of these first-time voters cast provisional ballots during the November 2004 election. To alert pollworkers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The city election office did not report experiencing challenges verifying the identification of these first-time voters. City of Madison, Wisconsin: In Madison, five provisional ballots were cast during the November 2004 election but none were counted. Provisional votes are counted when they are cast by qualified voters in their assigned precincts. The main reason these provisional votes were not counted was that the voters did not provide identification as required by HAVA for voters who applied by mail after January 2003 and were voting for the first time in the precinct. These voters were instructed to bring their identification to the local election office within 24 hours after the election but did not do so. The city election office did not report experiencing challenges implementing provisional voting. Voters were provided with written information at their polling places informing them that they could contact the local election office to find out the outcome of their provisional votes. However, the local election office did not provide information to voters on the outcome of provisional votes since this information is only provided when provisional votes count.[Footnote 37] The representative from the city election office did not know if first- time voters who registered by mail after January 2003 but did not send a copy of identification with their applications were allowed to cast provisional ballots if they were not able to provide identification at the polling place. To alert poll workers and elections staff about first-time voters who needed to provide identification at the polls because they did not provide it when they registered by mail, the voter registration list was marked with notes, codes, or marks next to the voters' names. The city election office reported experiencing challenges verifying the identification of these first-time voters. Specifically, the city election office required additional staff to send notices to these voters to notify them that they needed to submit the necessary documentation or bring it to the polling place on Election Day. [End of section] Appendix VIII: Reported Experiences of Selected Motor Vehicle Agency Offices with Voter Registration Applications: This appendix describes the responses of state and local motor vehicle agency (MVA) representatives to our survey about their experiences during 2004 in carrying out various voter registration application tasks--specifically, assisting MVA clients with completing voter registration applications, accepting applications from these clients, and forwarding applications to state or local elections offices. It also describes actions MVA officials reported that they took when individuals and state or local election officials contacted their offices about individuals who said they submitted their applications to MVA offices but their names were not on the voter registration lists when they went to vote in the November 2004 election. The statements in this appendix reflect the responses to our survey by the applicable state and local MVA officials and were not independently verified by us. In 1993, NVRA, also known as the "motor voter" law, was enacted. NVRA, among other things, established procedures to facilitate the registration of eligible citizens to vote in federal elections by providing them the opportunity to apply to register to vote at various locations, such as motor vehicle and other NVRA agencies primarily in states that have a voter registration requirement.[Footnote 38] Under NVRA, an eligible citizen conducting certain transactions at a motor vehicle agency (MVA)--such as applying for, renewing, or changing the address on a driver's license--can simultaneously apply to register to vote or, if applicable, change the voter's address. In 2001, we reported that about 46 percent of the jurisdictions responding to our nationwide survey expressed concerns about applications received from NVRA agencies. For example, officials most frequently noted challenges with processing incomplete or illegible applications, applications that arrived late at the local election office, and applications that never arrived. Because local election officials around the country expressed concerns about processing applications submitted at MVAs, as noted in our 2001 report, we surveyed MVA officials in six states[Footnote 39] and 12 local jurisdictions about their experiences during 2004 in carrying out voter registration application tasks and assisting voters who said they had submitted their applications to a MVA office but their names were missing from the voter registration lists when they went to vote in the November 2004 election. Appendix IX provides a copy of the survey sent to MVA officials in each of the six states and 12 local jurisdictions. Experiences of Selected State MVA Offices: As figure 8 illustrates, state MVA representatives reported, for the most part, that during 2004 MVA staff found that helping clients complete their voter registration applications, accepting applications from clients, and forwarding applications to state or local elections offices was either very easy, somewhat easy, or neither difficult nor easy. Figure 8: MVA Officials' Characterization of How Difficult or Easy It Was for MVA Staff to Assist Clients with Completing Voter Registration Applications, Accept Applications from Clients, and Forward Applications to Election Offices: [See PDF for image] [A] The state MVA officials responded not applicable when asked how difficult or easy it was for staff to help MVA clients complete the voter registration application because they said MVA staff do not help clients complete the application. [End of figure] Discussed below, by state, are the reasons MVA representatives reported that MVA staff did not experience difficulty performing the voter registration application tasks and the assistance staff provided to individuals who said they had submitted their applications to MVA offices, but their names were not on the voter registration lists when they went to vote in the November 2004 election. Arizona: When asked how difficult or easy it was for staff to help MVA clients complete the voter registration application, the state MVA representative responded "not applicable" because staff do not help MVA clients complete their voter registration applications. Accepting voter registration applications was easy because the electronic form walks the clients through each part of it. MVA clients cannot proceed through the form if all the required boxes have not been completed. Forwarding the applications to the state or local elections offices was more time consuming than difficult or easy. Twelve individuals contacted the state MVA office saying that they had submitted a voter registration application to their MVA office or another MVA office, but their names were missing from the voter registration lists when they went to vote in the November 2004 election. The MVA office referred these individuals to the local election office and checked the MVA electronic transaction information for their voter registration applications. State or local election officials contacted the state MVA office about individuals who said they were told their names were not on the voter registration lists when they went to vote in the November 2004 election. However, the state or local election officials did not know whether the individuals had said that they had submitted their voter registration applications to the state MVA office or another MVA office. California: When asked how difficult or easy it was for staff to help MVA clients complete the voter registration application, the state MVA representative responded not applicable because staff do not help MVA clients complete their voter registration applications. Accepting applications from MVA clients and forwarding these applications to state or local elections offices was very easy. The state MVA office was not contacted by individuals who said that their names were missing from the voter registration lists when they went to vote in the November 2004 election. But the MVA office was contacted by state or local elections officials who indicated that the individuals said they had submitted their voter registration applications to the state MVA office or another MVA office, but their names were missing from the voter registration lists when they went to vote in the November 2004 election. The state MVA representative, who did not know how many individuals had contacted the state or local elections officials, reported that when the elections officials provided adequate information, the MVA office researched the driver's license database to determine if the individuals' claims could be confirmed and informed the elections officials of the outcome of the MVA research. Michigan: Helping MVA clients to complete their voter registration applications and accepting the completed applications from them were very easy because the MVA offices process thousands of voter registrations annually following standard procedures that have not changed substantially in many years. Also, the voter registration applications are printed automatically with the voter name, address, and date of birth. The clients only have to verify the information in the application and sign it. Forwarding applications to the state or local elections offices was somewhat easy because clients are not always sure about their city or township. MVA offices are provided with pre-printed mailing labels for forwarding the applications to the various election jurisdictions, which makes the process quite simple overall. The state MVA representative did not know how many individuals or state or local elections officials had contacted the state MVA office or another MVA office about individuals who said that their names were not on the voter registration lists when they went to vote in the November 2004 election because this information was not collected or tracked. New York: The state MVA representative responded that helping MVA clients to complete their voter registration applications, accepting the completed applications from them, and forwarding the applications to state or local elections offices were neither difficult nor easy for MVA staff during 2004. The state MVA office has not been made aware of any difficulties that MVA staff might have encountered in carrying out these voter registration application tasks. Individuals and state or local elections officials had contacted the state MVA office about 421 individuals, in total, who said that they had submitted their voter registration applications to the state MVA office or another MVA office, but their names were not on the voter registration lists when they went to vote in the November 2004 election. The state MVA office searched its electronic files to determine whether the 421 individuals had indicated they wanted to apply to register to vote at the time of the MVA transaction. Also, if necessary, the state MVA office pulled the original source documents to determine the individuals' intent to apply to register to vote. Texas: The state MVA representative responded that helping MVA clients to complete their voter registration applications was neither difficult nor easy for MVA staff during 2004, but this task added time to processing MVA transactions. Accepting the completed applications from the clients also was neither difficult nor easy and added time to processing MVA transactions. MVA staff had to review the completed applications before accepting them and forwarding them to the state or local elections offices. Forwarding the completed voter registration applications to state or local election offices was neither difficult nor easy. Applications were either mailed or hand delivered to the state or local elections offices. Individuals and state or local elections officials had contacted the state MVA office about individuals who said that they had submitted their voter registration applications to the state MVA office or another MVA office, but their names were not on the voter registration lists when they went to vote in the November 2004 election. However, the state MVA office did not know how many individuals had said that they had submitted their voter registration applications to the state MVA office or another MVA office. Virginia: Helping MVA clients to complete their voter registration applications and accepting the completed applications from them were neither difficult nor easy because MVA has been handling voter registration since 1996. MVA staff provide the same level of service for voter registration as they do for other MVA transactions. Forwarding applications to the state or local elections offices is very easy because the applications are mailed daily to the state board of elections by all of the MVA offices. Four individuals contacted the state MVA office saying that they had submitted a voter registration application to their MVA office or another MVA office, but their names were missing from the voter registration list when they went to vote in the November 2004 election. The MVA office(s) and the state board of elections office researched various automated systems and paper records to determine when the clients had conducted business with the MVA and the disposition of any voter registration applications. The MVA clients were advised by letter of the outcome of this research. Also, the state board of elections office contacted the state MVA office indicating that 1,288 individuals had said that they had submitted their voter registration applications to the MVA state office or another MVA office, but their names were missing from the voter registration lists when they went to vote in the November 2004 election. MVA assisted the state board of elections office in verifying the individuals' claims and determining when they submitted an application to a MVA office. The state board of elections office, which has access to MVA records and houses all voter registration information, found that 295 individuals had applied to register to vote at a MVA office and their applications were sent to the state board of elections office. Of the remaining 993 individuals, 591 had indicated that they did not wish to apply to register to vote and did not complete an application; 245 had no record of conducting MVA business or submitting a voter registration application at a MVA office; 128 conducted an Internet, mail, or phone transaction with MVA but did not request that a paper voter registration application be mailed to them; and 29 submitted their voter registration applications after the registration deadline. Experiences of Selected Local MVA Offices: Gila County, Arizona: The MVA representative responded not applicable when asked how difficult or easy was it for staff to help MVA clients complete the voter registration application because staff do not help MVA clients complete their voter registration applications. The customer completes the application and MVA staff glance at it to ensure all the information has been completed. Accepting applications from clients and forwarding the applications to the elections offices is very easy because, after the clients complete their applications, a copy of the application is given to them and the original applications are mailed daily to the elections offices. The MVA office was not contacted by individuals or state/local election officials about individuals who said that their names were missing from the voter registration lists when they went to vote in the November 2004 election. Maricopa County, Arizona: When asked how difficult or easy it was for staff to help MVA clients complete the voter registration application, the MVA representative responded not applicable. Accepting applications from MVA clients and forwarding the applications to election offices was very easy. One individual contacted the MVA office saying that he/she had submitted a voter registration application to their MVA office or another MVA office, but his/her name was missing from the voter registration list when he/she went to vote in the November 2004 election. The MVA office referred the individual to the local election office. No state or local election officials contacted the MVA office about individuals who said they were told that their names were not on the lists of registered voters when they went to vote in the November 2004 election. Los Angeles County, California: The MVA representative responded not applicable when asked how difficult or easy was it for staff to help MVA clients complete the voter registration application because staff do not help MVA clients complete their voter registration applications. Accepting applications from clients and forwarding the applications to the elections offices is very easy because the applications are collected and mailed to the elections offices daily. The MVA office was not contacted by individuals or state/local election officials about individuals who said their names were missing from the voter registration lists when they went to vote in the November 2004 election. Yolo County, California: When asked how difficult or easy it was for staff to help MVA clients complete the voter registration application, the MVA representative responded not applicable because MVA staff do not assist MVA clients with completing voter registration forms. MVA staff provide clients the forms, accept the completed forms from the clients and note in the MVA database that the forms were received, and forward the forms to the local voter registrars. Accepting the forms from the clients and forwarding them to the elections offices was very easy. Most forms are forwarded weekly to the local registrars by mail and a special pickup is arranged for the deadline day before an election. The MVA office was not contacted by individuals or state/local election officials about individuals who said their names were missing from the voter registration lists when they went to vote in the November 2004 election. City of Detroit, Michigan: Assisting clients with completing their voter registration applications and accepting the applications from them was very easy because the applications are completed by computer as part of other transactions. The individuals only have to sign their applications. Forwarding applications to the elections offices was very easy because applications are presorted and pre-printed labels are used for mailing them to the elections offices weekly. Around the time of elections, applications are forwarded daily to the elections offices. The representative from the MVA office did not know whether individuals or state/local election officials had contacted their MVA office about individuals who said their names were missing from the voter registration lists when they went to vote in the November 2004 election because this information was not collected or tracked. Delta Township, Michigan: Assisting clients with completing their voter registration applications and accepting applications from them was very easy because thousands of these transactions are processed yearly; the clients check the information in their applications and sign them. Forwarding applications to the elections offices was neither easy nor difficult, but it can be time consuming based the volume of applications and the need to separate them by election jurisdictions. The representative from the MVA office did not know whether state or local election officials contacted the MVA office about individuals who said their names were missing from the voter registration lists when they went to vote in the November 2004 election because this information was not collected or tracked. Approximately 12 individuals contacted the MVA office saying that they had submitted their voter registration applications to their MVA office or another MVA office, but their names were missing from the voter registration lists when they went to vote in the November 2004 election. The MVA office referred the individuals to the local election office and tried to look up a record of the voter registration applications the individuals reportedly submitted. New York City, New York: Assisting MVA clients with completing the voter registration form, accepting forms from them, and forwarding the forms to the elections offices was very easy. When assisting clients with completing the forms, MVA staff question information on the voter registration forms only if the clients have not filled out any of the information on the voter registration forms. The MVA office was not contacted by individuals or state/local election officials about individuals who said that their names were missing from the voter registration lists when they went to vote in the November 2004 election. Rensselaer County, New York: Assisting MVA clients with completing their voter registration forms was very easy because MVA staff do not assist clients with completing the forms. Accepting the voter registration forms and forwarding them to the local elections offices was also very easy. The forms are mailed to the elections offices. Two individuals contacted the MVA office saying that they had submitted a voter registration application to their MVA office or another MVA office, but their names were missing from the voter registration lists when they went to vote in the November 2004 election. The MVA office referred the individuals to the local election office. No state or local election officials contacted the MVA office. Bexar County, Texas: Assisting MVA clients with completing the voter registration application, accepting applications from them, and forwarding the applications to the elections offices was very easy. The MVA office was not contacted by individuals or state/local election officials about individuals who said that their names were missing from the voter registration lists when they went to vote in the November 2004 election. Webb County, Texas: Assisting MVA clients with completing the voter registration application, accepting applications from them, and forwarding the applications to the elections offices is somewhat easy. The applications are very easy for the clients to complete, and staff forward the applications to the elections offices daily. MVA staff review incomplete applications for accuracy. The MVA office was contacted by individuals saying that their names were missing from the voter registration list when they went to vote in the November 2004 election. The representative from the MVA office did not know how many of these individuals contacted the MVA office. The MVA office was not contacted by state/local election officials about individuals who said they were told that their names were not on the lists of registered voters when they went to vote in the November 2004 election. Albemarle County, Virginia: Assisting MVA clients with completing their voter registration applications was neither difficult nor easy because MVA clerks are responsible for quickly reviewing the applications to ensure all parts have been completed and informing the clients of any parts of the application that are incomplete. If clients have questions about the voter registration application, MVA staff refer the clients to the local voter registrar or the state election office because the MVA is responsible for providing clients the opportunity to apply to register to vote, not answering voter registration questions or verifying clients' responses contained in the applications. Accepting applications from MVA clients was somewhat easy because sometimes the clients do not complete all parts of their applications and return incomplete applications to the MVA clerk. Forwarding applications to elections offices was also neither difficult nor easy because a policy and procedure are in place for carrying out this activity. The MVA office was not contacted by individuals or state/local election officials about individuals who said that their names were missing from the voter registration lists when they went to vote in the November 2004 election. Arlington County, Virginia: Assisting clients with completing their voter registration applications was neither difficult nor easy because MVA clerks do a quick review of the applications to make sure they are complete but do not verify any of the clients' responses. MVA offices are tasked with providing their clients with the opportunity to apply to register to voter. Accepting applications from the clients and forwarding them to the elections offices was somewhat easy. Forwarding these applications is a task that is performed daily. The MVA office was not contacted by individuals or state/local election officials about individuals who said that their names were missing from the voter registration lists when they went to vote in the November 2004 election. [End of section] Appendix IX: Survey of Motor Vehicle Agency Officials: [See PDF for images] [End of section] Appendix X: GAO Contact and Staff Acknowledgments: GAO Contact: William O. Jenkins, Jr. (202) 512-8777 or jenkinswo@gao.gov: Acknowledgments: In addition to the contact named above, John Mortin and Linda Watson, Assistant Directors; Leo Barbour; Amy Bernstein; Katherine Davis; Gina Flacco; Evan Gilman; Mary Martin; and Maria Santos made key contributions to this report. FOOTNOTES [1] Pub. L. No. 103-31, 107 Stat. 77 (1993). [2] Pub. L. No. 107-252, 116 Stat. 1666 (2002). With respect to provisional voting, under HAVA, states that had either (1) no voter registration requirements for voters with respect of federal elections (North Dakota) or (2) polling place registration on Election Day with respect to federal elections (Idaho, Minnesota, New Hampshire, Wisconsin, and Wyoming) in effect on and after August 1, 1994, are not subject to HAVA's provisional voting requirements. [3] Under HAVA, such individuals who desire to vote in person may cast a provisional ballot, whereas an individual desiring to vote by mail may cast a ballot that is to be treated as a provisional ballot. [4] States may refer to their motor vehicle agencies by different names. For purposes of this report, we will generally refer to them as "motor vehicle agencies." [5] GAO, Elections: Additional Data Could Help State and Local Election Officials Maintain Accurate Voter Registration Lists, GAO-05-478 (Washington, D.C.: June 10, 2005). [6] With respect to HAVA's statewide voter registration list requirements, while HAVA established a deadline of January 1, 2004, for states to have a statewide voter registration list and verification procedures, it also provided that states could request a waiver to extend the deadline to January 1, 2006. Eight states did not request a waiver, 42 states and the District of Columbia did request and received waivers, and 1 state (North Dakota) is not subject to HAVA's statewide voter registration list requirements. North Dakota is exempt from these requirements because it qualifies for a HAVA exemption applicable to any "State in which, under a State law in effect continuously on and after the date of enactment…there is no voter registration requirement for individuals in the State with respect to elections for Federal office." Pub. L. No. 107-252, 116 Stat. 1666, 1709 (2002). [7] GAO, Elections: Perspectives on Activities and Challenges across the Nation, GAO-02-3 (Washington, D.C.: Oct. 15, 2001). [8] We selected these states using a nonprobability sample--a sample in which some items in the population have no chance, or an unknown chance, of being selected. Results from nonprobability samples cannot be used to make inferences about a population. Thus, the information we obtained cannot be generalized to state and local election jurisdictions or MVA offices either nationwide or at the state level. [9] We defined a local jurisdiction with a large population as one having 150,000 or more people of voting age and a local jurisdiction with a small population as one having less than 150,000 people of voting age. The population figures are based on the 2000 U.S. Census. [10] We did not survey MVA officials in Wisconsin regarding voter registration under NVRA; MVA offices in Wisconsin do not participate in voter registration because the state is exempt from NVRA. Wisconsin's exemption from NVRA, in general, is based upon the fact that it did not, and still does not, have voter registration requirements for federal elections under state laws that were in effect on the NVRA- specified exemption date of August 1, 1994. [11] GAO, Elections: Additional Data Could Help State and Local Election Officials Maintain Accurate Voter Registration Lists, GAO-05- 478 (Washington, D.C.: June 10, 2005). [12] GAO, Elections: Perspectives on Activities and Challenges Across the Nation, GAO-02-3 (Washington, D.C.: Oct. 15, 2001). [13] While New York State had not passed legislation to implement provisional voting requirements in HAVA by the time of the November 2004 election, New York had a form of provisional voting in place referred to as affidavit ballots. The New York City election office said that letters were sent to voters whose affidavit ballots did not count. [14] Pub. L. No. 103-31, 101 Stat. 77 (1993). [15] States that had either (1) no voter registration requirement for voters with respect to federal elections (North Dakota) or (2) polling place registration on Election Day with respect to federal elections (Idaho, Minnesota, New Hampshire, Wisconsin, and Wyoming) in effect on and after August 1, 1994, are not subject to NVRA. [16] GAO, Elections: Perspectives on Activities and Challenges across the Nation, GAO-02-3 (Washington, D.C.: Oct. 15, 2001). [17] Federal Election Commission, The Impact of the National Voter Registration Act on Federal Elections 1999 - 2000. [18] Federal Election Commission, The Impact of the National Voter Registration Act on Federal Elections 2001 - 2002. [19] The National Commission on Federal Election Reform was organized by the Miller Center and The Century Foundation in early 2001 to formulate concrete proposals for election reform to "help ensure a more effective and fair democratic process in elections to come," as stated in the Commission's report, To Assure Pride and Confidence in the Electoral Process, August 2001. [20] The National Commission on Federal Election Reform's Task Force on the Federal Election System, July 2001. This task force report accompanied the report of the National Commission on Election Reform. [21] The National Commission on Federal Election Reform, To Assure Pride and Confidence in the Electoral Process, August 2001. [22] The Election Center is also known as the National Association of Election Officials and is composed almost exclusively of officials who serve in voter registration and administration of elections. [23] National Task Force on Election Reform, Election 2004: Review and Recommendations by the Nation's Election Administrators, May 2005. [24] Wisconsin and the two jurisdictions in Wisconsin were not included in our survey because, as noted in footnote 10, the state is exempt from NVRA. [25] GAO, Elections: Additional Data Could Help State and Local Election Officials Maintain Accurate Voter Registration Lists, GAO-05- 478 (Washington, D.C.: June 10, 2005). [26] HAVA, in general, requires states to, among other things, (a) implement an interactive statewide voter registration list for federal elections; (b) perform regular list maintenance by comparing the voter list against state records on felons and death; (c) match applicant information on voter registration lists with information in state motor vehicle agency's records; and (d) match voter registration application information on voter registration lists with Social Security Administration records, as appropriate. [27] GAO, Elections: Perspectives on Activities and Challenges Across the Nation, GAO-02-3 (Washington, D.C.: Oct. 15, 2001). [28] HAVA, in general, requires that provisional ballots be provided to first-time voters who registered to vote by mail on or after January 1, 2003, but were unable to show required identification when either voting in person or by mail in a federal election. [29] Local election officials in the two jurisdictions in Arizona said this was not applicable. In an April 2005 letter from the U.S. Department of Justice to the Arizona Secretary of State, Justice's Civil Rights Division concluded that it was permissible for a state to mandate that potential voters show identification at the polls prior to receiving provisional ballots. Local election officials in 2 of 14 jurisdictions surveyed reported that 12,815 first-time voters cast provisional ballots during the November 2004 election. The election officials representing the remaining jurisdictions said no first-time voters cast provisional ballots. [30] Total ballots cast does not necessarily represent the total number of votes cast and officially counted in any jurisdiction. In some jurisdictions, total ballots cast may include ballots where votes cast may or may not have been counted for a variety of reasons, including improperly marked ballots, ballots submitted without a mark for any particular candidate, ballots that included a mark for two candidates in the same election, or provisional ballots cast that were not counted. For purposes of this analysis, total ballots cast can include votes cast but not counted. For example, in Maricopa County, Arizona, the total number of ballots cast for president during the 2004 election was 1,211,963--which included 19,212 votes cast but not counted for a variety reasons--and the total number of votes cast and counted was 1,192,751. [31] While New York State had not passed legislation to implement provisional voting requirements in HAVA by the time of the November 2004 election, New York had a form of provisional voting in place referred to as affidavit ballots. The New York City election office said that letters were sent to voters whose affidavit ballots did not count. [32] GAO, Elections: Additional Data Could Help State and Local Election Officials Maintain Accurate Voter Registration Lists, GAO-05- 478 (Washington, D.C.: June 10, 2005) [33] We did not survey MVA officials in Wisconsin regarding voter registration under NVRA; MVA offices in Wisconsin do not participate in voter registration because the state is exempt from NVRA. Wisconsin's exemption from NVRA, in general, is based upon the fact that it did not, and still does not, have voter registration requirements for federal elections under state laws that were in effect on the NVRA- specified exemption date of August 1, 1994. [34] We did not send a survey to motor vehicle agency officials in Wisconsin because of its exemption from NVRA. [35] Local election officials in the two jurisdictions in Arizona said that the HAVA requirement to provide a provisional ballot in such instance was not applicable. In a letter from the U.S. Department of Justice to the Arizona Secretary of State, Justice's Civil Rights Division concluded that it was permissible for a state to mandate that potential voters show identification at the polls prior to receiving provisional ballots. [36] While New York State had not passed legislation to implement the provisional voting requirements in the Help America Vote Act by the time of the November 2004 election, New York had a form of provisional voting in place referred to as affidavit ballots. [37] With respect to provisional voting, under HAVA, states that had either (1) no voter registration requirements for voters with respect of federal elections (North Dakota) or (2) polling place registration on Election Day with respect to federal elections (Idaho, Minnesota, New Hampshire, Wisconsin, and Wyoming) in effect on and after August 1, 1994, are not subject to HAVA's provisional voting requirements. [38] Under NVRA, states that had either (1) no voter registration requirements for voters with respect to federal elections (North Dakota) or (2) polling place registration on Election Day with respect to federal elections (Idaho, Minnesota, New Hampshire, Wisconsin, and Wyoming) in effect on and after August 1, 1994, are not subject to NVRA. [39] Wisconsin was not included in our survey because MVA offices in this state do not participate in voter registration because, as noted in footnote 1 in this appendix, the state is exempt from NVRA. GAO's Mission: The Government Accountability Office, the investigative arm of Congress, exists to support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO's commitment to good government is reflected in its core values of accountability, integrity, and reliability. Obtaining Copies of GAO Reports and Testimony: The fastest and easiest way to obtain copies of GAO documents at no cost is through the Internet. GAO's Web site ( www.gao.gov ) contains abstracts and full-text files of current reports and testimony and an expanding archive of older products. The Web site features a search engine to help you locate documents using key words and phrases. You can print these documents in their entirety, including charts and other graphics. Each day, GAO issues a list of newly released reports, testimony, and correspondence. GAO posts this list, known as "Today's Reports," on its Web site daily. The list contains links to the full-text document files. To have GAO e-mail this list to you every afternoon, go to www.gao.gov and select "Subscribe to e-mail alerts" under the "Order GAO Products" heading. Order by Mail or Phone: The first copy of each printed report is free. Additional copies are $2 each. A check or money order should be made out to the Superintendent of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a single address are discounted 25 percent. Orders should be sent to: U.S. Government Accountability Office 441 G Street NW, Room LM Washington, D.C. 20548: To order by Phone: Voice: (202) 512-6000: TDD: (202) 512-2537: Fax: (202) 512-6061: To Report Fraud, Waste, and Abuse in Federal Programs: Contact: Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov Automated answering system: (800) 424-5454 or (202) 512-7470: Public Affairs: Jeff Nelligan, managing director, NelliganJ@gao.gov (202) 512-4800 U.S. Government Accountability Office, 441 G Street NW, Room 7149 Washington, D.C. 20548: