NSF awards $10 million for American National Election Studies
The National Science Foundation (NSF) has awarded $10 million to fund the American National Election Studies to study voter participation and decision-making in the 2012 U.S. presidential election, and in the mid-term elections of 2010. Read the press release.
Measuring unconscious attitudes
An important innovation in the 2008 studies was the incorporation of new measurement methods from psychology to assess unconscious racial prejudice. Read a report on the rationale for the Affect Misattribution Paradigm and some findings it generated.
The impact of calling cell phones on political surveys
The number of Americans who cannot be reached by a land line telephone at home and can only be reached by a cellular phone has been increasing. As a result, researchers conducting telephone surveys have begun to call cell phone numbers. In 2008, we included an innovative set of questions designed to identify respondents who are land line only, cell phone only, land line and cell phone, or unreachable by telephone at home. A report comparing answers to political questions given by these groups will be available in the coming months.
Turnout validation
Scholars have been increasingly interested in comparing survey reports of voter registration and turnout to official government records of registration and turnout for the same election. We are in the midst of gathering data from the States in order to make such comparisons possible for the 2008 surveys. A report will be posted here in the coming months.
Evaluating Representativeness of the 2008 Time Series and Panel Samples
The ANES surveys entailed interviewing probability samples of American adult citizens. To evaluate the success of the studies in yielding representative samples, we have compared the survey results to benchmarks to assess accuracy. A report will be posted here in the coming months.
Overstating Turnout in the ANES and other surveys
Although evaluations of probability sample survey data have often shown them to provide very accurate results, surveys during the last four decades have routinely overestimated the proportion of Americans who voted in an election, sometimes by a substantial margin. We have undertaken an investigation of such over-statement in the ANES as well as in the General Social Survey and the Current Population Survey to identify methodological factors that might contribute to it. A report will be posted here in the coming months.
Attrition in the 2008 ANES Panel
The 2008 panel study entailed 21 monthly waves of data collection from panel of respondents. We have undertaken an evaluation of the degree of respondent drop-out during the panel and assessed the extent to which this drop-out caused bias in the sample. A report will be posted here in the coming months.
Clarification of the Party Identification Summary variable
A small problem has been discovered with the coding of party identification in the ANES Time Series files. A recent review of prior study documentation revealed that the measure of party ID in ANES studies has been computed in slightly different ways in different years, and this coding been based partly on answers to questions about topics other than party identification. The inconsistency affects a very small proportion of respondents who have been assigned to the "apolitical" category. In most years, about 1 to 2 percent of respondents have been classified in ways that might surprise researchers and do not allow for appropriate comparisons over time. We have written a memo describing how party identification was constructed in each year  , so that users can be aware of the inconsistency and can make more informed decisions about how they prefer to code party identification in the future.
|
|
Advance Release of the ANES waves of the 2008-2009 Panel Study
An Advance Release of the 2008-2009 Panel Study dataset and documentation is now available for download. This release includes the data from the six ANES waves completed in 2008 as well as data from the recruitment and profile surveys.
The Advance Release allows interested users to analyze 2008 political data as soon as possible after the election. The data and documentation will be revised during 2009 prior to the full release of the study.
New Technical Report
A new technical report on How to Analyze ANES Survey Data, by Matthew DeBell is now available. This document provides guidance on how to access and analyze ANES data.
Coding answers to open-ended questions in the 2008 surveys
ANES sponsored a conference in December 2008 that brought together experts from many social science disciplines to discuss best practices in the coding of open-ended answers to survey questions. Read transcripts of the conference presentations.
For the first time, ANES has released all interviewer transcriptions of respondent answers to the open-ended questions in the 2008 surveys. Download the ZIP file.
An advisory committee was formed to propose new principles for coding open-ended measures of knowledge about politics. A report will be posted here in the coming months.
Measuring racial attitudes confidentially
The 2008 ANES involved an experiment to compare the measurement of racial stereotypes in two modes: direct oral reporting to an interviewer and confidential reporting using audio computer-assisted self-interviewing. The hypothesis explored is whether confidential reporting allows respondents to reveal prejudice that they would be reluctant to express to an interviewer. A report will be posted here in the coming months.
Discrepancies between ANES and GSS
During the last 35 years, the ANES and the General Social Survey have asked some identical questions repeatedly. We are in the process of comparing the results, documenting discrepancies in variable distributions and trends over time, and exploring explanations for the discrepancies to help analysts study trends more effectively. A report will be posted here in the coming months.
Question Wording Experiments in the 2008 Time Series
Most of the "core" ANES items were designed many decades ago, and the items vary considerably in format. During the ensuing years, the literature on questionnaire design has identified best practices for producing maximally valid and reliable measurements. The 2008 Time Series included many experiments in which traditional questions were asked of half the respondents, and improved questions measuring the same constructs were asked of the other half. A report comparing the results of the two sets of questions and a guide combining the two question versions for coordinated analysis will be available in the coming months. Read the memo explaining the Question Wording Experiments .
ANES Book of Ideas
In order to document the process by which the 2008 study questionnaires were designed, many scholars whose whose questions were evaluated in the 2006 pilot study wrote reports of analysis of the pilot study data. Their reports will appear in a forthcoming book. Read the table of contents .
|
