What is the Election Administration and Voting Survey?
The Election Administration and Voting Survey (also known as the Election Day Survey) provides information related to election administration, registration, and voting.

The survey includes national, state, and county-level data on:

• Voter Registration 
• Uniformed and Overseas Voters 
• Early, Absentee, and Provisional Voting 
• Voting Equipment Usage 
• Poll Workers, Polling Places, and Precincts

Who responds to the Election Administration and Voting Survey?
The Election Administration and Voting Survey is sent to election officials in 50 states, the District of Columbia, and four territories – Guam, Puerto Rico, American Samoa, and the U.S. Virgin Islands.

Why does the EAC produce the Election Administration and Voting Survey?
The EAC released its first survey in 2004 under authority granted to the commission by the Help America Vote Act (HAVA) of 2002. Section 202 of HAVA requires the EAC to serve as a national clearinghouse and resource for the compilation of information and review of procedures with respect to the administration of Federal elections. Section 202(3) authorizes the EAC to conduct studies and carry out other duties and activities to promote the effective administration of Federal elections.

HAVA mandates that the Commission collect information related to the processes and procedures used to register voters and to serve uniformed and overseas citizens wishing to vote. The National Voter Registration Act (NVRA) of 1993 required the Federal Election Commission (and subsequently, the Election Assistance Commission) to report to Congress by June 30 of the year following a Federal election on the impact of the Act on the administration of elections and to include recommendations for improvements in procedures, forms, and other matters affected by the Act.

Section 703 of HAVA mandates that for each regularly scheduled general election for Federal office, the EAC shall collect comprehensive data from the states on all of the ballots sent to military and overseas voters and received back by election administrators. The Uniformed and Overseas Citizens Absentee Voting Act (UOCAVA) of 1986 protects the voting rights of members of the uniformed services and U.S. citizens residing outside of the country.

In addition, the EAC uses the Election Administration and Voting Survey to learn more about other timely and important election administration issues, for example, from the use of electronic poll books to the prevalence of voter registration processes being conducted via the Internet.

How often does the EAC produce the Election Administration and Voting Survey?
The information collection associated with the Election Administration and Voting Survey is conducted every two years following a federal election.

Where can I read or download copies of the Election Administration and Voting Survey reports?
Three reports are produced from the data collected through the Election Administration and Voting Survey: 

• The Impact of the National Voter Registration Act of 1993 (NVRA) on the Administration of Elections for Federal Office
• Uniformed and Overseas Citizens Absentee Voting Act (UOCAVA) Survey Report Findings  
• Election Administration and Voting Survey (this comprehensive report includes key findings from the NVRA and UOCAVA reports, as well as findings from the other topical areas addressed in the survey)

The survey reports, including the data, for 2004, 2006, and 2008 are available to the public and can be found here.

What’s new in the 2008 Election Administration and Voting Survey?
New to the 2008 survey is a Statutory Overview in which states were asked a series of questions about their state election laws, definitions, and procedures. This information will help the EAC and its stakeholders understand the quantitative data appearing in the Election Administration and Voting Survey report, while also providing a clearer picture of the legal landscape governing U.S. elections. The Overview is available here.

The broad categories of questions to which states have been asked to respond are basically the same in 2008 as they were in 2004 and/or 2006 (i.e. Voter Registration, Uniformed and Overseas Voters, Early, Absentee, and Provisional Voting, Voting Equipment, and Poll Workers, Polling Places, and Precincts).

New questions appearing in the 2008 survey include, among others, voter identification for first-time voters, poll books, online registration, state laws regarding post-election audits and poll worker training. The 2008 survey instruments can be found here.

***

SURVEY METHODOLOGY AND APPROACH

How are EAC’s Election Administration and Voting Survey data collected?
Every two years, EAC administers a survey to 55 States and territories requesting election administration-related data at the county-level (or township, county equivalent). The 2008 survey includes data for the November 4, 2008 general election and the 2-year period leading up to it. The EAC reports the data it receives directly from the states.

Readers will sometimes observe complete jurisdictional level data for one question, but limited data reported for a different question. Election data collection provided by the 55 states and territories varied significantly. Most states relied at least to some degree on centralized voter registration databases (VRDs) and voter history da¬tabases, which allowed state election officials to respond to the survey at the local level for each question. Other states relied on cooperation from county election offices to complete the survey. Some states did not collect data in all the categories requested in the survey, and others did not have data for all their local jurisdictions for all question items. A jurisdiction is a generic term used to signify geographic areas that administer elections. The jurisdictions in this study may include counties, parishes, independent cities, towns or cities, or an entire state (e.g., Alaska).

Caution is necessary when interpreting the data, particularly when comparing the data from year to year or state to state, due to changes in survey question wording from year to year, state data collection prac¬tices across time, and the varying levels of completeness in states’ responses.

What is the methodology for the EAC survey?
EAC distributes the survey instrument to every state and territory. The data presented in the report were provided directly to EAC by those states and territories that participated, which is the same methodology EAC employed in its previous surveys. This approach differs from a sampling method, in which data are collected from a pre-determined segment deemed to be representative of an entire group. 

How many jurisdictions are there in the U.S.?
If every county/county equivalent, small city, town, and municipality (which are common to the
Midwest and New England regions) are included the number of jurisdictions is estimated to range from 7,000 to 10,000 depending on the how the reporting unit is defined.

The EAC asks states to provide survey data at the county or county-equivalent level. This produced a jurisdictional count of 4,517. This count is very close to the jurisdictional count that the Associated Press (AP) reported for 2008 (4,600).
 
States varied in the extent to which they reported data to the EAC at the county-level. While many states were able to report data for all of their jurisdictions, others were not and reported as a single unit, or reported only at the county level and not at the municipal level (such as New York, Illinois, Wisconsin, and Puerto Rico). This helps to explain some of the variation in jurisdictional counts.

There are varying levels of state responses within the same question. Why?
States were asked to provide data at the county (or county-equivalent/township level). Some states reported statewide totals, either because they consider themselves to be one jurisdiction (such as Alaska) or because of incomplete data reporting at lower levels (e.g., municipalities). Other states were able to report for all of their counties. Still, there are multiple sources of incomplete data (states missing counties, counties missing sub-jurisdictions), which is explained in the Election Administration and Voting Survey reports.

Although state data collection and reporting of the survey data for 2008 improved from the previous years, the data are still incomplete; readers should keep this in mind when examining the data contained in the reports.

Table 7 of the UOCAVA report, for example, shows that nearly all of the responding jurisdictions (4,427 out of 4,517) were able to report on the number of ballots transmitted, but only 2,475 out of 4,517 were able to report whether those ballots were one-time requests or automatic two-cycle requests (2,608 out of 4,517).

Similarly, Table 1 of the UOCAVA report summarizes how many of the 4,517 ju¬risdictions surveyed were able to provide data for selected questions. For example, when one sees that 91% of the jurisdictions were able to provide data on the number of UOCAVA ballots transmitted to overseas citizens, it means that for that question only, 4,131 out of 4,517 of the jurisdictions were able to provide data. It is not meant to suggest that the response rate for the entire survey was 91%

Information about the number of jurisdictions responding to each question and sub-question can be found in the tables in the appendix of the reports.

What information does the Election Administration and Voting Survey provide about voter participation?
The Election Administration and Voting Survey report provides the state-reported figures for U.S. voter participation and registration, as well as the estimated voting age population and estimated citizen voting age population as reported by the U.S. Census Bureau. Readers are encouraged to consult Table 30 to review these data.

What about a UOCAVA voter participation rate?
The 2008 UOCAVA report shows state and county-level data for the number of UOCAVA ballots transmitted, the number that were returned and submitted for counting, the number that were counted, and the number that were rejected. The data are also sorted by state and whether the voter was a military or overseas citizen voter.

Studies that attempt to calculate voter participation rates generally do so through sample surveys of individual voters. The data in the EAC’s Election Administration and Voting Survey were provided by the states, not individual UOCAVA voters. The survey was conducted by gathering raw data from the states and territories, not by using a sampling method as might be the case with national opinion surveys or surveys that attempt to determine participation rates of a subset of the population. For data collected from individual UOCAVA voters, read Voting from Abroad: A Survey of UOCAVA Voters. This research study surveyed UOCAVA voters regarding their reactions to, attitudes toward, and experiences with electronic voting.

There seem to be differences between the 2006 and 2008 UOCAVA data. For example, the number of UOCAVA ballots requested/transmitted seems to have changed between the two years. Why?

UOCAVA voters may receive their ballots in several ways. UOCAVA provides for the automatic transmission of ballots for two election cycles to voters who register an overseas address using the Federal Post Card Application (FPCA) (42 U.S.C. § 1973ff-3). Overseas voters may also request a ballot by applying to their local election of¬fice before a specific election.

The 2008 UOCAVA section of the survey differed from the 2006 version. New questions in 2008 included requests for information about automatically transmitted UOCAVA ballots, including the disposition of those ballots. The EAC also asked for breakdowns of both the type of ballots used (absentee or FWAB) and by the type of UOCAVA voter (military or overseas citizen).

A significant portion of the UOCAVA ballots transmitted are due to the automatic transmission provisions, which means that UOCAVA voters who registered two election cycles ago would have received their ballots automatically. It could be that some states experienced drop-off in the automatic ballots transmitted if the transmission of automatic ballots expired in 2008 for some voters, thereby helping to explain some of the change in the number of transmittals in 2006 versus 2008.

Also, the 2006 survey asked for data on ballots *requested*. While in theory all ballots requested should lead to a ballot *transmitted*, that may not always be the case.  For example, it is not known how many people who were automatically transmitted a ballot may have also made a request via another method.  It could also be the case that some states track ballot requests and ballot transmissions differently, either from a different source (voter history vs. voter registration roll) or they track one data item more closely than the other.

Lastly, it is difficult to gauge the number of domestic uniformed services voters who do not use the FPCA. While these voters are technically UOCAVA voters, election officials may not be able to classify them as such.

All of the aforementioned reasons could impact the “number of ballots transmitted” figure. EAC suggests that readers exercise caution when attempting to compare the 2006 and 2008 data and not draw inferences beyond the scope of the data collected.

What do zero, N/A, blanks, not categorized, and unknown mean as used in the Election Administration and Voting Survey reports?States have a number of options for reporting data for the various questions and sub-questions. The 2008 survey instrument instructed states to enter “N/A” if data were not available or if a question was not applicable to a state or county. If a state entered a zero, that response was treated as a valid response equaling zero (“0”). A “blank” response was treated as a nonresponse.
The Notes section for the tables, which appears in each report, explains the Balance/Not Categorized column. The Balance/Not Categorized column compares the sum of all the response columns with the total column. For example, in Table 40 of the Election Administration and Voting Survey report, the total number of poll workers is 878,360. The age categories listed in the columns adds up to 379,926, which means there is an uncharacterized balance of 498,434 poll workers whose ages are not known.

Within the table itself, if the number shown under the balance column is a positive number the difference is treated as an uncategorized response (meaning the respondent could not slot it into one of the categories listed). If the number shown under the balance column is a negative number (indicated by the parentheses) this indicates that the sum of the responses is greater than the total. This could occur by an error in data entry or by the inability to correctly categorize some responses, resulting in some over-counting. For example, in Table 40 of the same report, one state shows a balance with a negative number, meaning that the total number of poll workers reported (2,564) was less than the numbers reported in the columns (2,589). The difference, or over-counting, could have been due to a data entry error or an inability to slot some of the poll workers into the correct age categories. 

How can I learn more about a specific state’s data?
EAC makes all of its data publicly available for interested parties to examine and analyze. The Election Administration and Voting Survey data can be found on the EAC website. For questions about a specific state’s data, please review the data and contact the state’s election office for more information.

Questions about the survey?
Please call the EAC at 1-866-747-1471/202-566-3100 or send an e-mail to HAVAinfo@eac.gov