Why Are Surveys Struggling to Estimate Vote Shares (ICPSR 39673)
Version Date: Jun 10, 2026 View help for published
Principal Investigator(s): View help for Principal Investigator(s)
Matthew Tyler, Rice University;
D. Sunshine Hillygus, Duke University;
Ted Brader, University of Michigan;
Matthew DeBell, Stanford University;
Shanto Iyengar, Stanford University;
Daron R. Shaw, University of Texas;
Nicholas A. Valentino, University of Michigan
Series:
https://doi.org/10.3886/ICPSR39673.v1
Version V1
Summary View help for Summary
This release is a replication resource for the journal article "Why Are Surveys Struggling to Estimate Vote Shares?" (American Journal of Political Science, 28 April 2026). This collection contains data, codebooks, code, and additional documentation to replicate the assessment. The study centers around the question as to whether political surveys are accurate. It is in response to the 2020 U.S. presidential election, during which the political support of candidate Donald Trump was significantly underestimated. The study leverages a large set of data collected alongside the 2020 American National Election Studies (ANES) to identify and quantify sources of error using the Total Survey Error framework. This methodological approach quantifies a non-ignorable nonresponse bias, and reveals that the estimated differential nonresponse is sufficiently large to explain the ANES 2020 Biden-Trump error. This result has broader implications for the presence of bias in political surveys, and informs potential avenues of correction in future survey development.
Citation View help for Citation
Export Citation:
Subject Terms View help for Subject Terms
Geographic Coverage View help for Geographic Coverage
Smallest Geographic Unit View help for Smallest Geographic Unit
County
Restrictions View help for Restrictions
Access to these data is restricted. Users interested in obtaining these data must complete a Restricted Data Use Agreement, specify the reason for the request, and obtain IRB approval or notice of exemption for their research.
Distributor(s) View help for Distributor(s)
Time Period(s) View help for Time Period(s)
Data Collection Notes View help for Data Collection Notes
-
These data are a Fast Track Release and are distributed as they were received from the data depositor. The files have been zipped for release, but not checked or processed.
Study Purpose View help for Study Purpose
This study aims to reevaluate the accuracy of political surveys in the United States. It illuminates evidence of nonignorable unit nonresponse, specifically that supporters of Donald Trump during the 2020 U.S. presidential election were less likely to respond to political surveys. Trump supporters were also less likely to complete the 2020 American National Election Study (ANES) survey if they did respond. The overall goal of this study is to diagnose and quantify bias in the current political survey framework in order to inform corrections in future survey design.
Study Design View help for Study Design
This study implements the Total Survey Error framework, which offers an approach to understanding, measuring, and reducing errors in probability sample surveys. It refers to an accumulation of error that arises from the difference between true population characteristics and a survey's estimates of those characteristics. This use of TSE focuses on the four most plausible sources of error which explain the underestimation of Trump support: coverage error, unit nonresponse error, item-nonresponse error, and measurement error.
Sample View help for Sample
The data includes a 2016-2020 panel, voter files, and a nonresponse follow-up (NRFU) survey. The 2020 sample is made up of both panel and cross-section components. The panel component contains repeat interview data from the 2016 American National Election Study (ANES, sample size 2,839), and also sampled from the United States Postal Service Delivery Sequence File (DSF). A mixture of the ANES 2020 (fresh sample) respondents and nonrespondents were invited to a nonresponse follow-up (NRFU) study, data that is meant to help characterize the households that declined to participate in the ANES. The ANES target population contains all possible voters, and as such the data could be subsetted to make direct comparisons with pre-election polls. Subset comparison allows a Democrat-Republican margin (DRM) to be calculated.
Time Method View help for Time Method
Universe View help for Universe
Registered voters throughout counties in the United States
Unit(s) of Observation View help for Unit(s) of Observation
Data Source View help for Data Source
American National Election Studies. ANES 2020 Time Series Study Geocodes. Inter-university Consortium for Political and Social Research [distributor], 2024-03-28. https://doi.org/10.3886/ICPSR38176.v3
U.S. Census Bureau, "TIGER/Line Shapefile, 2019, 2010 nation, U.S., 2010 Census Urban Area National", https://catalog.data.gov/dataset/tiger-line-shapefile-2019-2010-nation-u-s-2010-census-urban-area-national
MIT Election Data and Science Lab, 2018, "County Presidential Election Returns 2000-2020", https://doi.org/10.7910/DVN/VOQCHQ, Harvard Dataverse, V13, UNF:6:GILlTHRWH0LbH2TItBsb2w== [fileUNF]
American National Election Studies. 2021. ANES 2020 Time Series Study Full Release [dataset and documentation]. February 10, 2022 version. www.electionstudies.org
Manson, Steven, Jonathan Schroeder, David Van Riper, Katherine Knowles, Tracy Kugler, Finn Roberts, and Steven Ruggles. 2024. IPUMS National Historical Geographic In- formation System: Version 19.0 [dataset]. Minneapolis, MN. https ://doi . org / http: //doi.org/10.18128/D050.V19.0.
American National Election Studies. 2017. ANES 2016 Time Series Study Full Release [dataset and documentation]. September 4, 2019 version. www.electionstudies.org
User's Guide and Codebook for the ANES 2020 Time Series Voter Validation Supplemental Data. 2022. Ann Arbor, MI and Palo Alto, CA: the University of Michigan and Stanford University. Accessed March 28, 2023. www.electionstudies.org.
American National Election Studies. 2021. ANES 2020 Non-Response Follow-Up Study [dataset and documentation]. October 28, 2021 version. Accessed April 3, 2023. www.electionstudies.org.
American National Election Studies. 2021. ANES 2020 Methodology Dataset Release [dataset and documentation]. July 19, 2021 version. Accessed April 3, 2023. www.electionstudies.org.
U.S. Census Bureau, "TIGER/Line Shapefile, 2019, nation, U.S., Current Metropolitan Statistical Area/Micropolitan Statistical Area (CBSA) National", https://catalog.data.gov/dataset/tiger-line-shapefile-2019-nation-u-s-current-metropolitan-statistical-area-micropolitan-statist
US Census Bureau. 2021. 2010 Census Urban and Rural Classification and Urban Area Criteria. https://www.census.gov/programs-surveys/geography/guidance/geo-areas/urban-rural/2010-urban-rural.html.
Data Type(s) View help for Data Type(s)
Mode of Data Collection View help for Mode of Data Collection
Description of Variables View help for Description of Variables
This evaluation approach uses survey data from the American National Election Studies (ANES). It references the 2020 ANES, which uses a probability-based design sampled from the United States Postal Service Delivery Sequence File (DSF), a list of residential addresses. The data was compiled through a process of vote validation comparison in order to assess whether false reporting of turnout is correlated with vote choice. The variables include demographic information about voters, such as age, sex, gender, marital status, income level, region of residence, and birth nation category. Other variables include voting decision in the 2016 and 2020 elections, as well as political beliefs. The variables also consist of geographical information, showing voter representation by county.
Response Rates View help for Response Rates
The 2020 ANES response rate was approximately 40%.
HideNotes
These data are freely available to data users at ICPSR member institutions. The curation and dissemination of this study are provided by the institutional members of ICPSR. How do I access ICPSR data if I am not at a member institution?
One or more files in this data collection have special restrictions. Restricted data files are not available for direct download from the website; click on the Restricted Data button to learn more.