AP VoteCast, Presidential Primaries, United States, 2024 (ICPSR 39139)

Version Date: Nov 4, 2024 View help for published

Principal Investigator(s): View help for Principal Investigator(s)
Jennifer Benz, Associated Press-NORC Center for Public Affairs Research; Trevor Tompson, Associated Press-NORC Center for Public Affairs Research

Series:

https://doi.org/10.3886/ICPSR39139.v1

Version V1

Slide tabs to view more

AP VoteCast is a survey of the American electorate conducted by NORC at the University of Chicago for The Associated Press and Fox News. The survey is funded by AP. In 2024, AP VoteCast covered presidential primary elections in three states across three election dates. For all three elections, interviews were conducted via phone and web. The surveys concluded when polls closed on Election Day: The Iowa Republican caucus on January 15, the New Hampshire Republican and Democratic primaries on January 23, and the South Carolina Republican primary on February 24. Respondents who completed the survey three days or more before Election Day (for example, January 8-12 for Iowa) were asked if they would be willing to be re-contacted for a follow-up survey. Those who consented were contacted during the final three days of the field period and re-asked their intent to vote and vote choice. The vote intent and vote choice provided in the re-contact survey were used for the final estimates. If a respondent did not complete the re-contact survey, their original intent to vote and vote choice responses were retained. Respondents were re-contacted using their preferred mode -- a call from a NORC interviewer or a text or email with a link to complete the short re-contact survey online.

AP VoteCast combines a random sample of registered voters drawn from state voter files with a sample of self-identified registered voters selected from nonprobability online panels. Interviews with the probability and nonprobability samples were conducted in English and Spanish. Participants in the probability sample selected from state voter files were contacted by phone and mail and had the opportunity to take the survey by phone or online. All interviews from nonprobability sample sources were completed online.

Benz, Jennifer, and Tompson, Trevor. AP VoteCast, Presidential Primaries, United States, 2024. Inter-university Consortium for Political and Social Research [distributor], 2024-11-04. https://doi.org/10.3886/ICPSR39139.v1

Export Citation:

  • RIS (generic format for RefWorks, EndNote, etc.)
  • EndNote
Associated Press

State

Inter-university Consortium for Political and Social Research
Hide

2024-01-08 -- 2024-02-24
2024-01-08 -- 2024-02-24
  1. For additional information, please visit the study website.
Hide

The purpose of this study is to survey the American electorate.

AP VoteCast combines a random sample of registered voters drawn from state voter files with a sample of self-identified registered voters selected from nonprobability online panels. Interviews with probability and nonprobability samples were conducted in English and Spanish. Respondents received a small monetary incentive for completing the survey. Participants in the probability sample selected from state voter files were contacted by phone and mail and had the opportunity to take the survey by phone or online. All interviews from nonprobability sample sources were completed online.

Prior to dialing, all probability sample records were mailed a postcard inviting them to complete the survey either online using a unique PIN or via telephone by calling a toll-free number. Postcards were addressed by name to the sampled registered voter if that individual is under age 35; postcards were addressed to "[STATE] Registered Voter" in all other cases. For sampled voters who did not respond online, additional outbound dialing was conducted for sampled records based on their predicted response propensity. Telephone interviews were conducted with the adult who answered the phone. Both online and telephone respondents provided confirmation of registered voter status in the state. Interviews were conducted in English and Spanish.

Nonprobability participants were provided by Dynata, Cint, and Prodege, including members of their third-party panels. Dynata used router technology to recruit participants, and all available panelists age 18 and older in each state were recruited. Dynata's system used built-in technology that uses digital fingerprinting, geolocation clues, and checks at enrollment to confirm identity and to identify suspicious behavior to prevent respondents from completing the survey more than once. Cint's suppliers invited respondents to the survey using email invites and panelist recruitment. Before sending them into the survey, Cint targeted and pre-screened respondents age 18 and older on the basis of registered voter status and state location using zip codes. Prodege invited its members to the survey using invitations sent by email and through the panelist portal. To prevent respondents from completing more than once, Prodege verified panelists using email double opt-in verification, physical address verification, device fingerprinting, mobile verifications, and CAPTCHA while continuously monitoring their users to avoid panelists duplication. From there, they managed how many surveys were sent to each user, and only allowed each user to enter the survey once. Beyond each sample provider's protocols described above, NORC also applies a digital fingerprint software and panel-level ID validation to prevent respondents from completing the VoteCast survey multiple times. Nonprobability respondents provided confirmation of their registered voter status in the state.

Probability-based Registered Voter Sample: In each of the three states, NORC obtained a sample of registered voters from Catalist LLC's registered voter database. This database included demographic information, addresses, and phone numbers for registered voters, allowing potential respondents to be contacted via mail and telephone. Noncoverage is limited to registered voters who register to vote immediately prior to the election (e.g., states with same day registration laws) and are therefore not on the voter rolls at the time of sampling. The sample was stratified by state, partisanship (either party affiliation on the voter file or the Catalist vote choice index variable), past primary participation, age, and race. In addition, NORC attempted to match sampled records to a registered voter database maintained by L2, which provided additional phone numbers. After the matching, NORC had phone numbers for 90% of sampled records, including cell phone numbers for 87% of records with a phone number.

Nonprobability Sample: Nonprobability participants were provided by Dynata, Cint, and Prodege, including members of their third-party panels. In Iowa, among the 595 panelists who touched the Dynata's pre-screener instrument, 315 went on to complete the full survey. In New Hampshire, among the 790 panelists who touched the pre-screener instrument, 494 went on to complete the full survey. In South Carolina, among the 652 panelists who touched the pre-screener instrument, 211 went on to complete the full survey. Panelists recruited for a specific state were only allowed to complete the survey if they said they were registered to vote in that state. In Iowa, among the 945 panelists who touched the Cint's pre-screener instrument, 328 went on to complete the full survey. In New Hampshire, among the 1,070 cases who touched the pre-screener instrument, 156 went on to complete the full survey. In South Carolina, among the 1,099 panelists who touched the pre-screener instrument, 158 went on to complete the full survey. Respondents recruited for a specific state were only allowed to complete the survey if they said they were registered to vote in that state. In order to ensure the final sample did not include any respondents who completed the survey more than once, Cint removed duplicates by participant ID and cookies. Before sending them into the survey, Prodege targeted and pre-screened respondents age 18 and older on the basis of age, gender, education, income, race/ethnicity, state, and registered voter status. In Iowa, among the 1,345 panelists who touched the Prodege's pre-screener instrument, 298 went on to complete the full survey. Respondents recruited for a specific state were only allowed to complete the survey if they said they were registered to vote in that state. In New Hampshire, among the 755 panelists who touched the pre-screener instrument, 337 went on to complete the full survey. In South Carolina, among the 1,470 panelists who touched the pre-screener instrument, 353 went on to complete the full survey.

A response rate cannot be calculated for nonprobability samples. While there is no way to quantify the size of the non-covered population for an opt-in panel, the primary population least likely to be included was those without internet access.

In Iowa, 2,658 registered voters completed the survey, including 1,717 probability interviews from state voter files (1,499 online and 218 via telephone) and 941 nonprobability interviews completed online. The margin of sampling error was plus or minus 3.4 percentage points for voters (N=1,597) and 5.2 percentage points for nonvoters (N=1,061).

In New Hampshire, 3,487 registered voters completed the survey, including 2,500 probability interviews from state voter files (2,061 online and 439 via telephone) and 987 nonprobability interviews completed online. The margin of sampling error was plus or minus 3.5 percentage points for voters (N=1,989) and 12.3 percentage points for nonvoters (N=185) for the Republican primary. The margin of sampling error was plus or minus 4.6 percentage points for voters (N=1,093) and 11.4 percentage points for nonvoters (N=220) for the Democratic primary.

In South Carolina, 3,122 registered voters completed the survey, including 2,400 probability interviews from state voter files (2,126 online and 274 via telephone) and 722 nonprobability interviews completed online. The margin of sampling error was plus or minus 3.4 percentage points for voters (N=2,466) and 6.5 percentage points for nonvoters (N=656).

All registered voters in Iowa and South Carolina, and registered voters registered as Democrats, Republicans or unaffiliated in New Hampshire.

Individuals

  • The overall weighted response rate for the probability sample drawn from the Iowa voter files was 2.2%.
  • The overall weighted response rate for the probability sample drawn from the New Hampshire voter files was 3.2%.
  • The overall weighted response rate for the probability sample drawn from the South Carolina voter files was 2.7%.

Hide

2024-11-04

2024-11-04 ICPSR data undergo a confidentiality review and are altered when necessary to limit the risk of disclosure. ICPSR also routinely creates ready-to-go data files along with setups in the major statistical software formats as well as standard codebooks to accompany the data. In addition to these procedures, ICPSR performed the following processing steps for this data collection:

  • Created variable labels and/or value labels.
  • Standardized missing values.
  • Created online analysis version with question text.

Hide

VoteCast employed a four-step weighting approach for each of the three state surveys that combines the probability sample with the nonprobability sample and refined estimates at a subregional level within each state.

Hide

Notes