American National Election Studies, Panel Study, 2008-2009

DOI

10.17605/OSF.IO/Y3UKX

Citation

Krosnick, J. A., Lupia, A., & Hutchings, V. (2020, May 5). American National Election Studies, Panel Study, 2008-2009.

Summary

The 2008-2009 ANES Panel Study is a series of surveys of a representative sample of the American electorate recruited by telephone. Panelists began completing monthly surveys on the Internet in January 2008. The study is intended to support research on candidate choice and voter turnout in the 2008 presidential election. The chief purpose of the advance release is to make interim data available to the user community as quickly as possible after the presidential election. The advance release includes all survey responses to questions about the election that were asked in 2008, except for a small amount of data that has been redacted because its release could pose a risk to respondent privacy.

To minimize panel attrition and conditioning effects, only seven of the twenty-one monthly surveys are about politics. Other surveys are about a variety of non-political topics. The panelists answered political questions prepared by ANES in January, February, June, September, October and November 2008. With certainty, the panel will answer more political questions in May 2009. It is also possible that panelists will answer a limited number of political questions on other 2009 waves. The advance release includes data from the six ANES-created political surveys of 2008, as well as the recruitment and profile surveys. The full release will include data from all 21 waves. Note that the 2008-2009 ANES Panel Study is entirely separate from the 2008 ANES Time Series study, which was conducted using the traditional ANES method of face-to-face interviews before and after the 2008 election. Although there are a few questions common to both studies, the samples and methods are different.

For more information, see the User's Guide.

The ARDA has added five additional variables to the original data set to enhance the users' experience on our site.

Data File

Cases: 3049
Variables: 2960
Weight Variable: WGTCR, WGTCSP, WGTCS01, WGTCS02, WGTCS06, WGTCS09, WGTCS10, WGTCS11, WGTC02, WGTC06, WGTC09, WGTC10, WGTC11, WGTL09, WGTL10, WGTL11

Data Collection

January 2008 through November 2008

Funded By

This material is based upon work supported by the National Science Foundation under Grant Nos. SES-0535332 and SES-0535334. ANES is also supported by Stanford University and the University of Michigan.

Collection Procedures

Profile Survey

After completing the recruitment interview on the telephone (in 97.2 percent of cases) or online (in 2.8 percent of cases), panelists were asked to complete a profile survey online. The profile survey served two purposes. First, it introduced panelists to the online survey format and trained them to answer survey questions on the Internet, which was expected
to be a novel experience for many, perhaps most, of the panelists. Next, the profile asked demographic background questions about the panelists. The profile survey began with introductory questions about reasons for joining the panel
and any difficulties the panelist may have experienced. These questions introduced respondents to questions of the following formats: multiple-choice questions allowing one response, multiple-choice questions allowing multiple responses, questions with pulldown menus to choose responses, and questions that ask the respondent to type words
for their answer.

Monthly Survey Procedures

Panelists were invited to complete one survey each month on the Internet. Invitations to these surveys were sent by e-mail. If the respondent did not complete the survey within several days, reminder e-mails were sent. Telephone prompting was initiated when repeated e-mail invitations did not yield a response.

Sampling Procedures

The Panel Study is designed to represent the population of U.S. citizens age 18 and older as of November 4, 2008 (Election Day). The sample for the Panel Study consisted of two cohorts. The first cohort, recruited in late 2007, consisted of 12,809 landline telephone numbers. The second cohort, recruited in the summer of 2008, consisted of 10,720 landline telephone numbers, for a total of 23,529 telephone numbers in the two cohorts combined. Knowledge Networks (KN) called each of these numbers to attempt to recruit an eligible person to participate in the Panel Study. A person was eligible if he or she was 1) a U.S. citizen, 2) born on or before November 4, 1990, and 3) residing in a household served by a sampled landline telephone number at the time of recruitment.

Recruitment spanned a period of four months for the first cohort and about 3.5 months for the second. No formal limit on the number of callbacks existed and in practice numbers were called up to 50 times to attempt an interview. At each telephone number, an interviewer attempted to identify all eligible household members, randomly select one eligible person for recruitment to the Panel Study, and complete a recruitment interview with the sampled person to invite that person to join the Panel Study. In all communications with prospective participants, the study was referred to by using a title appropriate for a periodic survey on a variety of topics, in order to avoid nonresponse bias that could be associated with interest in politics if the name "American National Election Study" were used. To prevent respondents from associating the study with ANES, the exact title used in respondent communications is confidential.

For more information on sampling, see the User's Guide.

Principal Investigators

Jon A. Krosnick, Arthur Lupia, and Vincent Hutchings

Weights

The weights on the advance release file have not been subject to complete review by ANES and are preliminary and subject to revision. The weighting documentation in this section is intended to provide a general description of these preliminary weights. More comprehensive documentation of the weights will accompany the full release of the Panel Study data in late 2009. Weights are designed to make the data more representative of the population of interest, namely people eligible to vote in the 2008 U.S. presidential election (operationally defined as U.S. citizens age 18 or older as of November 4, 2008). To make the data more representative, the weights count respondents in proportion to the size of the population group that each individual respondent represents. Hence, weights are a means of counting some respondents more than others when using the data to reliably describe the population of interest. The weights on the advance release file were calculated by Knowledge Networks following procedures specified by ANES.

Cross-sectional weight

The cross-sectional weight is non-zero for any respondent who completed a given wave of the study. It is zero for each respondent who did not complete a given wave. This weight is calculated for all eight stages of the study on the Advance Release data file. For example, the cross-sectional weight for wave 9 is non-zero for each respondent who completed wave 9 and is zero for everyone else who did not complete wave 9. The cross-sectional weight is for an analysis of data from just one wave. Note: The cross-sectional weight may not be optimal for analysis of a given wave and prior waves together
because the panelists who responded to a prior wave may not be the same panelists who responded to the current wave. As a result, cases may be unnecessarily dropped from an analysis, or the weight used may have been calculated for optimal use with a larger number of cases. For such analyses, a cumulative weight may be more appropriate.

Cumulative ANES panel weight

This weight is non-zero for respondents who completed a given ANES wave and all prior ANES waves. For example, the cumulative ANES panel weight for wave 9 is non-zero for each respondent who completed waves 1, 2, 6, and 9, and is zero for everyone else.

Cumulative late panel weight

Cohort 2 added new respondents to the panel beginning with wave 9, marking a new phase in the study. This weight is nonzero for respondents who completed a given wave and all prior waves from wave 9 onward. For example, this weight for wave 11 is non-zero for respondents who completed waves 9, 10, and 11, whether those respondents are part of cohort 1 or cohort 2.