Data sources
Every dataset the book draws on, with its years, weights and license, and the chapters that use it.
ANES 2024 Time Series
American National Election Studies 2024 Time Series Study
- Id
- anes2024
- Producer
- American National Election Studies (Stanford University and the University of Michigan)
- Unit
- respondent (5,521 rows; 4,964 completed the post-election interview)
- Years
- 2024 (pre-election and post-election interviews)
- Weights
- V240107b for post-election items (PSU V240107c, stratum V240107d); V240107a for pre-election items. Mode-specific weights: V240101a and V240101b (fresh in-person sample), V240102a and V240102b (fresh web sample).
- License
- Free download after accepting the ANES terms of use; cite the release.
Cite as: American National Election Studies. ANES 2024 Time Series Study [dataset]. Stanford University and the University of Michigan. https://electionstudies.org.
Holds the 2024-only items the cumulative file lacks: 2020 election legitimacy, January 6, political violence, DEI, transgender policy, Ukraine and Israel aid, boycotting and buycotting, trust in news media, and the usable 2024 religion items. The 2,416 respondents in neither the fresh in-person nor the fresh web sample are the 2016-2020-2024 panel and paper cases. The user guide names the full-sample weights and the design variables used here. The file also carries the 2020 and 2016 case identifiers of panel respondents, but the earlier studies are not yet used in the book.
Used in The two coalitions, Immigration, Guns, Trust in government (planned), Political violence and democratic norms, Education and ideology (planned), Religious tradition and race.
ANES Cumulative Data File
American National Election Studies Time Series Cumulative Data File, 1948-2024
- Id
- anes
- Producer
- American National Election Studies (Stanford University and the University of Michigan)
- Unit
- respondent (73,745 rows in the harmonized file)
- Years
- 1948-2024 (33 study years; even years only after 2004; no 2006, 2010 or 2014 study)
- Weights
- weight (VCF0009z). For face-to-face and web estimates use weight_ftf and weight_web, which exist for 2008, 2012, 2016 and 2024 and not for 2020.
- License
- Free download after accepting the ANES terms of use; cite the release.
Cite as: American National Election Studies. ANES Time Series Cumulative Data File, 1948-2024 [dataset]. Stanford University and the University of Michigan. https://electionstudies.org.
Geography is state only, so no county claim can rest on ANES. Interviewing was mainly face-to-face through 2008, then mixed with web from 2012; 2020 and 2024 are mixed mode. The 2020 study had no in-person interviewing, and 2020 has no mode weights. In 2024 the mode weights cover only the fresh samples (the fresh in-person and fresh web samples), not the panel and paper cases. In the 2024 rows of this file the religion and church attendance coding is unusable (the Protestant count falls while Catholic is flat, and attendance is missing for about half of respondents), so 2024 religion comes from the 2024 Time Series or the CES. Several long-run items are retired in 2024, including the trust in government index. Year-specific items, such as the 2024 boycott question, are in the separate Time Series release (dataset anes2024). Self-reported vote leans slightly Democratic.
Used in Party, belief and the political map, Party and ideology, Sort dates by attitude, Cohort change in partisan sorting, Constraint among issue positions (planned), Political knowledge and constraint (planned), The South (planned), Feelings toward the other party, Negative partisanship, Affective and issue polarization (planned), Party and feelings toward social groups (planned), The two coalitions, The diploma divide, Hispanic, Black and Asian voters (planned), The gender gap in attitudes and votes, Age and generation, Turnout, The abortion sort, Immigration, Guns, Perceptions of the economy, Health care, Trade and economic populism (planned), Foreign policy (planned), Fifty years of racial attitudes, Racial attitudes and ideology, Racial attitudes after George Floyd's death (planned), Policing (planned), Trust in government (planned), Generalized and institutional trust (planned), The conservative label (planned), Economic and cultural dimensions, Class, church and the two axes (planned), The obedience-autonomy disposition, The god gap in America, Religious tradition and race, Two moralities (planned), Down the ballot, Who donates (planned), Party and place.
ANES panel studies
ANES panel studies, including the 2016-2020-2024 panel component
- Id
- anes-panel
- Producer
- American National Election Studies (Stanford University and the University of Michigan)
- Unit
- respondent
- Years
- 2016 to 2024.
- Weights
- Panel weights; the 2024 Time Series names V240106a and V240106b for its panel sample.
- License
- Free download after accepting the ANES terms of use.
Cite as: American National Election Studies. ANES panel study files. https://electionstudies.org. Record each release at acquisition.
The separate panel files are not yet used in the book. The 2024 Time Series has a panel sample (sample type 1, 2,171 respondents interviewed in 2016, 2020 and 2024) and the 2016 and 2020 case identifiers, but the earlier studies are needed to see the same respondents' earlier answers.
Used in Movers and opinion change (planned).
Area Deprivation Index
Neighborhood Atlas Area Deprivation Index
- Id
- adi
- Producer
- University of Wisconsin School of Medicine and Public Health
- Unit
- census block group (the atlas files carry a national rank and a state decile)
- Years
- Vintage as distributed with the tract and ZCTA atlas files.
- Weights
- none.
- License
- The Neighborhood Atlas may restrict commercial use and redistribution; read its terms before publishing derived tract values.
Cite as: Kind, Amy J. H., and William R. Buckingham. 2018. Making neighborhood-disadvantage metrics accessible: the Neighborhood Atlas. New England Journal of Medicine 378: 2456-2458.
Used in the book as adi_national_rank and adi_state_decile in the tract and ZCTA atlas files. Higher ranks mean more deprivation.
Used in Measures of hardship compared, Deprivation and mobility (planned).
CDC COVID-19 county data
CDC county-level COVID-19 deaths and vaccination
- Id
- cdc-covid
- Producer
- Centers for Disease Control and Prevention
- Unit
- county
- Years
- Not yet acquired.
- Weights
- none.
- License
- See the provider's terms of use.
Cite as: Centers for Disease Control and Prevention. COVID-19 county files on data.cdc.gov. Record the dataset names and download date at acquisition.
Not yet used in the book. Needed for the COVID and vote chapter. State reporting gaps in county vaccination counts are expected and must be checked at acquisition.
Used in COVID and the vote (planned).
CDC PLACES
PLACES: Local Data for Better Health
- Id
- cdc-places
- Producer
- Centers for Disease Control and Prevention
- Unit
- census tract and ZCTA (model-based estimates of adult prevalence)
- Years
- Release as distributed with the tract and ZCTA atlas files.
- Weights
- none.
- License
- US government data, open access.
Cite as: Centers for Disease Control and Prevention. PLACES: Local Data for Better Health. https://www.cdc.gov/places/. Record the release year at citation.
Used in the book as about 40 health_* columns in the tract and ZCTA atlas files. The values are model-based (small-area estimates built from survey data and census composition), so they are not an independent check on census-derived deprivation measures and their uncertainty is not carried in the files. For ZCTAs the health_source column marks values as native, aggregated from tracts, or mixed.
Used in Measures of hardship compared, Health outcomes and the vote (planned).
CDC WONDER
CDC WONDER county-level mortality files
- Id
- cdc-wonder
- Producer
- Centers for Disease Control and Prevention, National Center for Health Statistics
- Unit
- county by cause of death and year
- Years
- Not acquired; the edition and years are to be recorded at acquisition.
- Weights
- none (death counts and rates).
- License
- US government data.
Cite as: Centers for Disease Control and Prevention, National Center for Health Statistics. Multiple Cause of Death files on CDC WONDER. Record the edition and query date at acquisition.
Not yet used in the book. The tract atlas carries small-area injury mortality columns (inj_*: firearm, homicide, suicide and overdose rates) whose source is not documented, so the book does not describe them as CDC WONDER output.
Used in Deaths of despair (planned).
CES panel files
Cooperative Election Study panel components
- Id
- ces-panel
- Producer
- Harvard University, fielded online by YouGov
- Unit
- respondent
- Years
- Panel waves across several election cycles; not yet acquired.
- Weights
- Panel weights supplied with each file.
- License
- See the provider's terms of use.
Cite as: Cooperative Election Study panel files. Harvard Dataverse. Record each file DOI at acquisition.
Not yet used in the book. The cumulative common content is repeated cross-sections, so it cannot show individual change. Panel attrition needs its own weights.
Used in Movers and opinion change (planned).
Cooperative Election Study
Cooperative Election Study (formerly the Cooperative Congressional Election Study), cumulative common content, 2006-2024
- Id
- ces
- Producer
- Harvard University, fielded online by YouGov
- Unit
- respondent (701,955 rows)
- Years
- 2006-2024 (every year; self-reported turnout is available in the 10 even years)
- Weights
- weight_cumulative for pooled years; weight for a single wave; vvweight (validated vote) exists for 2018-2024 only.
- License
- See the provider's terms of use.
Cite as: Cumulative CES Common Content. Harvard Dataverse, doi:10.7910/DVN/II2DB6.
The only survey used here with county FIPS. The weights are post-stratification weights for an opt-in panel, so intervals are model-based. The policy battery ends in 2021. Alaska county columns are missing by construction, 2006 and 2007 carry 2004 returns as the county environment, and county cells need at least 50 respondents. The file includes the 2026-07-12 repair of the county context for 2024 respondents in South Carolina, Wisconsin and West Virginia. The clean file has no validated turnout and no House, Senate or Governor vote choice; both exist in the upstream cumulative file.
Used in Party, belief and the political map, Party and ideology, Sort dates by attitude, Constraint among issue positions (planned), Affective and issue polarization (planned), The two coalitions, The diploma divide, The income ladder, Hispanic, Black and Asian voters (planned), The gender gap in attitudes and votes, Age and generation, Immigration, Guns, Perceptions of the economy, Health care, Trade and economic populism (planned), Climate and the environment (planned), Fifty years of racial attitudes, Policing (planned), The conservative label (planned), Economic and cultural dimensions, The god gap in America, The rise of the nones (planned), Religious tradition and race, Income and the county vote (planned), Education, ethnicity and the county shift, 2016 to 2024, Deprivation and mobility (planned), Stated and funded ideology (planned), Party and place, Immigration and local context (planned), Variation in opinion within places (planned), Partisans in the other party's territory (planned).
County demographics panel
County demographics for presidential election years, 2000-2024
- Id
- county-demographics
- Producer
- US Census Bureau (decennial census and American Community Survey) and USDA Economic Research Service (Rural-Urban Continuum Codes), compiled for this project
- Unit
- county
- Years
- 2000-2024 (election years, interpolated between census and survey anchors)
- Weights
- none (full counts and interpolated estimates).
- License
- US government source data, public domain. The interpolated panel is derived work.
Cite as: US Census Bureau. Decennial census and American Community Survey. USDA Economic Research Service. Rural-Urban Continuum Codes, 2013 and 2023 editions.
Values between anchor years are linear interpolations (the demog_*_src columns record provenance). Percent columns are proportions from 0 to 1. Income uses the CPI-deflated real 2024 series. The voting-age population column is not the voting-eligible population and overstates the denominator where many residents are non-citizens. The 2013 rural-urban edition is used through 2016 and the 2023 edition for 2020 and 2024.
Used in Reconstruction and the modern map (planned), Five axes of the county vote, The diploma map, Income and the county vote (planned), Rural counties (planned), Suburbs and density (planned), Education, ethnicity and the county shift, 2016 to 2024, The geography of donations (planned), Counties where donors and voters diverge, Immigration and local context (planned), Composition and behavior in county change (planned).
County election returns
US county election returns: President 1868-2024, House and Senate 1990-2024
- Id
- county-returns
- Producer
- Carlos Algara and Sharif Amlani (President 1868-2020); MIT Election Data and Science Lab (President 2000-2024); Paul Heideman (House and Senate, 1990-2024)
- Unit
- county or county equivalent
- Years
- President 1868-2024 (40 elections); House and Senate 1990-2024
- Weights
- none (full counts). Vote-weight county averages by two-party votes and report the county-equal version where the two differ.
- License
- The Heideman file is CC BY 4.0 (doi:10.5281/zenodo.22976553), but the state records behind it keep their own terms, and the sites of Florida, Indiana, Colorado and Wisconsin restrict reuse.
Cite as: Algara, Carlos, and Sharif Amlani. County-level presidential returns, 1868-2020. Harvard Dataverse, doi:10.7910/DVN/DGUMFI. MIT Election Data and Science Lab. County presidential returns. Harvard Dataverse. Heideman, Paul (2026). U.S. county-level election results: President, U.S. House and U.S. Senate, 1990-2024, version 1.0.0. Zenodo, doi:10.5281/zenodo.22976553.
Use the 1868-2024 and 2000-2024 files, not the legacy 1972-2024 file, whose 2024 totals are doubled for 173 counties in South Carolina, Wisconsin and West Virginia. Kansas City, Missouri is county 2938000 in every year of the modern file. Alaska reports by election district (absent before 2024) and is dropped from cross-year work. Not a balanced panel: 50 defunct counties have no FIPS. Third-party years make the two-party leader differ from the plurality winner. House and Senate returns cover the 48 contiguous states, with 41 state-year-offices not fully covered.
Used in Party, belief and the political map, The South (planned), Negative partisanship, Nationalization of the county vote, Distance from the national vote, 1868 to 2024, Down the ballot, The Electoral College (planned), Reconstruction and the modern map (planned), Five axes of the county vote, The diploma map, Income and the county vote (planned), Rural counties (planned), Education, ethnicity and the county shift, 2016 to 2024, The geography of donations (planned), Counties where donors and voters diverge, Party and place, Partisans in the other party's territory (planned), Composition and behavior in county change (planned).
CPS Voting and Registration Supplement
Current Population Survey, November Voting and Registration Supplement, 1976-2024, harmonized by IPUMS CPS (Version 13.0)
- Id
- cps
- Producer
- US Census Bureau, with the Bureau of Labor Statistics; harmonized and distributed by IPUMS CPS (University of Minnesota)
- Unit
- person (3,350,644 rows in the cleaned file)
- Years
- 1976-2024 (25 November samples, biennial, even years)
- Weights
- VOSUPPWT, the Voter Supplement weight, for every analysis of the supplement variables. The cleaned file carries no replicate weights.
- License
- IPUMS CPS conditions of use: a free extract after registration, no fee may be charged for the data, IPUMS must be cited, and publications are to be added to the IPUMS bibliography. The underlying Census Bureau data are public domain.
Cite as: Sarah Flood, Miriam King, Renae Rodgers, Steven Ruggles, J. Robert Warren, Daniel Backman, Etienne Breton, Grace Cooper, Julia A. Rivera Drew, Stephanie Richards, David Van Riper, and Kari C.W. Williams. IPUMS CPS: Version 13.0 [dataset]. Minneapolis, MN: IPUMS, 2025. https://doi.org/10.18128/D030.V13.0
On disk as the cleaned file cps_voter.parquet, built from IPUMS extract cps_00017: 25 November samples from 1976 to 2024 and 3,350,644 person rows. Not yet used in a chapter. Turnout and registration are self-reported, and the self-reported turnout rate exceeds the official count of ballots cast. Turnout coded by the Census convention reproduces the published Census series for 2008 to 2024. Citizenship is recorded from 1994. County is identified for about 300 counties from 1996 (232 to 355 a year), so county estimates cover only the identified counties. Family income (FAMINC) is in brackets that are not harmonized across years. The mode of voting (VOTEHOW) and reason for not voting (VOWHYNOT) items start in 1996 and are unvalidated. The file has no replicate weights, so design-based intervals cannot be computed from it.
Used in The two coalitions, Turnout.
DIME
Database on Ideology, Money in Politics, and Elections, public version 4.0
- Id
- dime
- Producer
- Adam Bonica, Stanford University
- Unit
- donor (44.2 million records) and candidate or committee (479,502 records)
- Years
- 1979-2024 (23 election cycles)
- Weights
- none (administrative records, no sampling).
- License
- ODC-BY 1.0, as listed on the Stanford Libraries page (read through a search result, not the license text).
Cite as: Bonica, Adam. 2024. Database on Ideology, Money in Politics, and Elections: Public version 4.0. Stanford University Libraries. https://data.stanford.edu/dime.
A donor's CFscore is fixed, so a change over time is a change in who gives, not in any donor's views. Donor geography and occupation are the most recent record. The donor file does not say which candidates a donor gave to. Itemized giving after 2008 reflects ActBlue and WinRed reporting. Candidate records are federal only, and 2024 winners and vote shares are not coded in this release. Use the reliable-donor flag for headline figures and show all donors as a check.
Used in Polarization in Congress (planned), Who donates (planned), The geography of donations (planned), Occupation and donations (planned), The ideology of the donor pool, Candidate extremity and vote share (planned), Stated and funded ideology (planned), Counties where donors and voters diverge, Partisans in the other party's territory (planned).
EJScreen
EPA Environmental Justice Screening and Mapping Tool
- Id
- ejscreen
- Producer
- US Environmental Protection Agency
- Unit
- census block group (the atlas files carry tract values)
- Years
- Version as distributed with the tract atlas file.
- Weights
- none.
- License
- US government data. The EPA removed EJScreen from its website in February 2025, so the archived copy is the source to cite.
Cite as: US Environmental Protection Agency. EJScreen: Environmental Justice Screening and Mapping Tool. Archived copy maintained by Public Environmental Data Partners, https://screening-tools.com/epa-ejscreen.
Used in the book as the ej_* columns in the tract atlas (fine particles, ozone, nitrogen dioxide, diesel particulate matter, air toxics, traffic proximity, and proximity to Superfund, risk management and hazardous waste sites).
Used in Measures of hardship compared, Air pollution and the vote.
European Social Survey
European Social Survey, rounds 1 to 11 cumulative
- Id
- ess
- Producer
- ESS ERIC (European Research Infrastructure Consortium)
- Unit
- respondent (540,671 rows in the clean file)
- Years
- 2002-2023 (rounds 1 to 11; round 11 was fielded in 2023 and 2024)
- Weights
- weight (= anweight = pspwght x pweight), correct for pooling countries.
- License
- Open access after free registration in the ESS Data Portal; follow the ESS data citation requirements.
Cite as: ESS ERIC. European Social Survey, rounds 1-11 [data files]. Distributed through the ESS Data Portal (Sikt).
The country set changes by round (21 countries in round 7, 31 in round 10), so trends fix the country set. Modules rotate: welfare items appear in rounds 4 and 8 only, democracy evaluations in rounds 6 and 10, climate items in rounds 8, 10 and 11, and the Schwartz value items in rounds 1 to 10. The clean file has no party vote; the raw Datafile.sav does. No design-based standard errors without the sampling design file.
Used in Climate and the environment (planned), Institutional trust outside the United States (planned), Support for democracy, Left and right across countries, Non-placement on the left-right scale (planned), Basic values and left-right, The silent revolution (planned), The obedience-autonomy disposition, Order, security and threat (planned), National pride and grievance (planned), The god gap in America, Secularization in Europe (planned), Immigration and local context (planned).
General Social Survey
General Social Survey, 1972-2024 cumulative
- Id
- gss
- Producer
- NORC at the University of Chicago, with support from the National Science Foundation
- Unit
- respondent (75,699 rows; the book uses a 161-column analysis subset and the 6,942-column raw file)
- Years
- 1972-2024 (35 study years)
- Weights
- weight_use (wtssnrps from 2004, wtssps before); design variables vstrat and vpsu.
- License
- See the provider's terms of use.
Cite as: General Social Survey, 1972-2024 cumulative data. NORC at the University of Chicago. https://gss.norc.org.
Geography is census region only. Interview mode is recorded from 2004 in the raw file, and the book joins it to the analysis subset. Waves through 2018 are mostly in person with some phone interviews; 2021 is about 87% web by respondent count, and 2022 and 2024 mix in-person and web interviews, so every post-2020 estimate needs a mode split or a mode note. Item coverage differs by year.
Used in Party, belief and the political map, Party and ideology, Sort dates by attitude, The South (planned), The diploma divide, The gender gap in attitudes and votes, The abortion sort, Immigration, Guns, Health care, Trade and economic populism (planned), Foreign policy (planned), Fifty years of racial attitudes, Racial attitudes after George Floyd's death (planned), Policing (planned), Confidence in thirteen institutions, Trust in science (planned), Generalized and institutional trust (planned), Tolerance of speech, The conservative label (planned), Class, church and the two axes (planned), The obedience-autonomy disposition, System justification (planned), The god gap in America, The rise of the nones (planned), Religious tradition and race, Two moralities (planned).
IRS SOI ZIP code data
IRS Statistics of Income, individual income tax statistics by ZIP code
- Id
- irs-soi
- Producer
- Internal Revenue Service, Statistics of Income Division
- Unit
- ZIP code (ZCTA in the atlas)
- Years
- Tax year as distributed with the ZCTA atlas file.
- Weights
- none.
- License
- See the provider's terms of use.
Cite as: Internal Revenue Service, Statistics of Income Division. Individual Income Tax Statistics, ZIP Code Data.
Used in the book as the irs_* columns in the ZCTA atlas (mean adjusted gross income per return, shares of returns with the earned income credit, business income, capital gains, itemized deductions, and the average tax rate). IRS county-to-county migration data are a separate product and are not yet used in the book.
Used in Wages, capital and the ZIP code vote, Measures of hardship compared.
Nationscape
Democracy Fund + UCLA Nationscape
- Id
- nationscape
- Producer
- Democracy Fund Voter Study Group and UCLA (Chris Tausanovitch and Lynn Vavreck)
- Unit
- respondent (494,532 rows)
- Years
- 2019-07-18 to 2021-04-15 (85 weekly waves)
- Weights
- weight. Not weight_2020, which covers a sparse subsample.
- License
- Free after registration and acceptance of the Voter Study Group terms; also on Harvard Dataverse.
Cite as: Tausanovitch, Chris, and Lynn Vavreck. 2021. Democracy Fund + UCLA Nationscape, full data set (version 20211215). Harvard Dataverse, doi:10.7910/DVN/CQFP3Z.
An opt-in online panel with post-stratification weights, so intervals are model-based and time series should cluster on week. The weekly cadence supports event designs, but COVID overlaps every 2020 event, many items rotate (check first_week and last_week in the dictionary), 38 rows lack a state, and the final weekly bins are thin.
Used in Sort dates by attitude, Affective and issue polarization (planned), The gender gap in attitudes and votes, Immigration, Guns, Trade and economic populism (planned), Climate and the environment (planned), Racial attitudes after George Floyd's death (planned), Policing (planned).
Opportunity Atlas
The Opportunity Atlas: tract-level outcomes of children by parental income
- Id
- opportunity-atlas
- Producer
- Opportunity Insights and the US Census Bureau
- Unit
- census tract
- Years
- Birth cohorts and outcome years as defined by Chetty and others (2018).
- Weights
- none.
- License
- See the provider's terms of use.
Cite as: Chetty, Raj, John N. Friedman, Nathaniel Hendren, Maggie R. Jones, and Sonya R. Porter. 2018. The Opportunity Atlas: Mapping the Childhood Roots of Social Mobility. NBER Working Paper 25147.
Used in the book as the kfr, kir, jail and teenbrth columns by race and sex in the tract atlas. These are model-based estimates for children who grew up in the tract, not for its current residents.
Used in Measures of hardship compared, Deprivation and mobility (planned).
Precinct presidential returns
Precinct-level presidential returns, 2016, 2020 and 2024
- Id
- precinct
- Producer
- MIT Election Data and Science Lab
- Unit
- precinct by candidate by voting mode
- Years
- 2016, 2020, 2024
- Weights
- none (full counts).
- License
- See the provider's terms of use.
Cite as: Baltz, Samuel, and others. 2022. American election results at the precinct level. Scientific Data 9: 651. doi:10.1038/s41597-022-01745-0. The 2024 release is cited from its Dataverse record.
Precinct units are redrawn every cycle, so precincts are not a panel; join years on county FIPS. Two-party totals need the mode-safe sum (a precinct total row where present, otherwise the sum of its mode rows). Use the candidate last name, not the simplified party label, for two-party work, because the 2016 file misses fusion lines. Votes hidden for privacy in 2024 are missing.
Used in Precincts within counties, Sorting by geographic scale, 2016 to 2024 (planned).
Precinct votes on census geographies
Precinct votes allocated to ZIP code areas, tracts and block groups by Regionalized Land Cover Regression, 2016 and 2020
- Id
- rlcr
- Producer
- Ali Fekrazad, using VEST (2016) and MIT Election Data and Science Lab (2020) precinct returns
- Unit
- ZCTA, census tract, block group
- Years
- 2016 and 2020 (contiguous United States only)
- Weights
- none. Vote counts are fractional; weight area-level statistics by total presidential votes and drop units with 100 votes or fewer.
- License
- CC0 (data doi:10.7910/DVN/Z8TSH3), as stated in the project README.
Cite as: Fekrazad, Amir. 2025. A dataset of US precinct votes allocated to Census geographies with precision. Scientific Data 12: 794. doi:10.1038/s41597-025-05140-3.
Ecological estimates: area-level, not individuals. The ZIP code area is the only geography that lines up between 2016 and 2020; tracts change vintage between the two years, so tract work is a 2020 cross-section. No Alaska or Hawaii. There is no 2024 release; a build path exists in the project scripts but has not been run.
Used in Sorting by geographic scale, 2016 to 2024 (planned), Wages, capital and the ZIP code vote, Suburbs and density (planned), Measures of hardship compared, Firearm deaths and the vote (planned), Deprivation and mobility (planned), Health outcomes and the vote (planned), Social capital (planned), Air pollution and the vote.
State presidential returns
State-level presidential returns, 1976-2024
- Id
- state-returns
- Producer
- MIT Election Data and Science Lab
- Unit
- state and candidate
- Years
- 1976-2024 (13 elections)
- Weights
- none (full counts).
- License
- See the provider's terms of use.
Cite as: MIT Election Data and Science Lab. State presidential returns, 1976-2024. Harvard Dataverse. https://electionlab.mit.edu/data.
Useful for state history, candidate names and as a benchmark for survey vote choice. The electoral vote counts for 2020 and 2024 are not in this file; a separate workbook holds them for 1972-2016 only.
Used in The two coalitions, The diploma divide, Age and generation, Turnout, Religious tradition and race, Down the ballot, The Electoral College (planned).
Voteview
Voteview: Congressional Roll-Call Votes Database (DW-NOMINATE scores)
- Id
- voteview
- Producer
- Jeffrey B. Lewis, Keith Poole, Howard Rosenthal, Adam Boche, Aaron Rudkin and Luke Sonnet
- Unit
- member of Congress by Congress
- Years
- 1st to 119th Congress (1789 to 2027)
- Weights
- none.
- License
- Free CSV download; cite as requested. The page states no formal license text.
Cite as: Lewis, Jeffrey B., Keith Poole, Howard Rosenthal, Adam Boche, Aaron Rudkin, and Luke Sonnet (2026). Voteview: Congressional Roll-Call Votes Database. https://voteview.com/
Not yet used in the book. DIME carries first-dimension DW-NOMINATE for candidate-cycle rows from 1980 to 2024 (dwnom1), which covers members but not the pre-1980 series, the second dimension or roll calls.
Used in Polarization in Congress (planned).
World Values Survey and EVS
World Values Survey and European Values Study, pooled waves 1 to 7
- Id
- wvs
- Producer
- World Values Survey Association and the European Values Study (GESIS Data Archive)
- Unit
- respondent (666,907 rows in 118 country codes)
- Years
- 1981-2023 (7 unified waves)
- Weights
- weight (S017) within a country; cross-country summaries compute each country first and give countries equal weight.
- License
- Free download with registration; users must cite the release.
Cite as: EVS/WVS (2024). European Values Study and World Values Survey: Joint EVS/WVS 2017-2022 Dataset. GESIS Data Archive, Cologne, ZA7505, version 5.0.0, doi:10.4232/1.14320. Earlier waves cite their own releases.
The country label is missing for ten EVS countries in the cleaned file (Austria, Belgium, Northern Cyprus, Denmark, Iceland, Ireland, Luxembourg, Malta, Portugal, Kosovo); the book restores them from the country code. Left-right self-placement does not mean the same thing in every country, so value-to-ideology work is done within country. Some items exist in only some waves. Country contexts such as the human development index are missing for a few places.
Used in Climate and the environment (planned), Institutional trust outside the United States (planned), Support for democracy, Left and right across countries, Non-placement on the left-right scale (planned), Education and ideology (planned), Basic values and left-right, Countries in value space (planned), The silent revolution (planned), The obedience-autonomy disposition, Order, security and threat (planned), National pride and grievance (planned), System justification (planned), The god gap in America, Secularization in Europe (planned), Two moralities (planned).