If you want to read more about how and why the Times/Siena Poll is conducted, you can see answers to frequently asked questions and submit your own questions here.
The New York Times/Siena College polls of 7,879 voters across the seven battleground states were conducted in English and Spanish on cellular and landline telephones from Oct. 24 to Nov. 2. The polls include 1,004 voters in Georgia and 1,010 voters in Nevada from Oct. 24 to Nov. 1, 1,025 voters in Arizona from Oct. 25 to Nov. 2, Oct. 28 to Nov. 2 voters in North Carolina from Oct. 28 to Nov. 2, 1,305 voters in Wisconsin from Oct. 25 to Nov. 2, and 1,527 voters in Pennsylvania and 998 voters in Michigan from Oct. 29 to Nov. 2, 2024.
For each poll, the margin of sampling error among the likely electorate is plus or minus 3.4 percentage points in Arizona, plus or minus 4.1 percentage points in Georgia, plus or minus 3.7 percentage points in Michigan, plus or minus 3.6 percentage points in North Carolina, plus or minus 3.6 percentage points in Nevada, plus or minus 2.9 percentage points in Pennsylvania, and plus or minus 3.4 percentage points in Wisconsin.
Sample
The survey is a response-rate-adjusted stratified sample of registered voters taken from the voter file maintained by L2, a nonpartisan voter-file vendor, and supplemented with additional voter-file-matched cellular telephone numbers from Marketing Systems Group. The sample was selected by The New York Times in multiple steps to account for differential telephone coverage, nonresponse and significant variation in the productivity of telephone numbers by state.
To adjust for noncoverage bias, the L2 voter file for each state was stratified by statehouse district, party, race, gender, marital status, household size, turnout history, age and homeownership. The proportion of registrants with a telephone number and the mean expected response rate were calculated for each stratum. The mean expected response rate was based on a model of unit nonresponse in prior Times/Siena surveys. The initial selection weight was equal to the reciprocal of a stratum's mean telephone coverage and modeled response rate. For respondents with multiple telephone numbers on the L2 file, or with differing numbers from L2 and Marketing Systems Group, the number with the highest modeled response rate was selected.
In Wisconsin, a state with a history of having a significant share of voters register on Election Day, a supplemental sample of unregistered voters was obtained from Marketing Systems Group. The New York Times excluded records who appeared to match with voters registered on the L2 voter file, based on their name, gender, age and location. The likely electorate for Wisconsin includes 304 interviews from this sample.
Fielding
The sample was stratified according to political party, race and region. Marketing Systems Group screened the sample to ensure that the cellular telephone numbers were active, and the Siena College Research Institute fielded the poll, with additional fieldwork by ReconMR, the Public Opinion Research Laboratory at the University of North Florida, the Institute of Policy and Opinion Research at Roanoke College, the Center for Public Opinion and Policy Research at Winthrop University in South Carolina, and the Survey Center at the University of New Hampshire. Interviewers asked for the person named on the voter file and ended the interview if the intended respondent was not available. Overall, more than 98 percent of respondents were reached on a cellular telephone.
The questions were translated into Spanish by ReconMR. Bilingual interviewers began the interview in English and were instructed to follow the lead of the respondent in determining whether to conduct the survey in English or Spanish. Monolingual Spanish-speaking respondents who were initially contacted by English-speaking interviewers were recontacted by Spanish-speaking interviewers. Overall, 12 percent of the interviews (14 percent of the weighted sample) among self-reported Hispanics were conducted in Spanish.
An interview was determined to be complete for the purposes of inclusion in the questions about whom the respondent would vote for if the respondent did not drop out of the survey after being asked the two self-reported variables used in weighting -- age and education -- and answered at least one of the questions about age, education or presidential-election candidate preference.
Weighting (registered voters)
The survey was weighted by The Times using the survey package in R in multiple steps.
First, the sample was adjusted for unequal probability of selection by stratum.
Second, each poll was weighted to match voter file-based parameters for the characteristics of registered voters.
The following targets were used:
* Party (party registration if available in the state, else classification based on participation in partisan primaries if available in the state, else classification based on a model of vote choice in prior Times/Siena polls) by whether the respondent's race is modeled as white or nonwhite (L2 model)
* Validated record of advance voting in 2024 (L2 data and data compiled by The New York Times), including by party outside of Wisconsin.
* Race or ethnicity (L2 model)
* Age (self-reported age, or voter-file age if the respondent refused) by gender (L2 data)
* Education (four categories of self-reported education level, weighted to match NYT-based targets derived from Times/Siena polls, census data and the L2 voter file)
* White/nonwhite race by college or noncollege educational attainment (L2 model of race weighted to match NYT-based targets derived from census data in states other than Georgia; L2 model of race weighted to match NYT-generated targets derived from census data in Georgia)
* Turnout history (NYT classifications based on L2 data)
* Method of voting in the 2020 elections (NYT classifications based on L2 data), in states other than Pennsylvania
* State region (NYT classifications)
* History of voting in the 2020 presidential primary (L2 data), in North Carolina
* Census tract educational attainment in Michigan (NYT classifications based on Census data).
Finally, the sample of respondents who completed all questions in the survey was weighted identically as well as to the result for the general-election horse-race question (including voters leaning a certain way) on the full sample.
Weighting (likely electorate)
The survey was weighted by The Times using the R survey package in multiple steps.
First, the samples were adjusted for unequal probability of selection by stratum.
Second, the first-stage weight was adjusted to account for the probability that a registrant would vote in the 2024 election, based on a model of turnout in the 2020 election.
Third, the sample was weighted to match targets for the composition of the likely electorate. The targets for the composition of the likely electorate were derived by aggregating the individual-level turnout estimates described in the previous step for registrants on the L2 voter file. The categories used in weighting were the same as those previously mentioned for registered voters.
Fourth, the initial likely electorate weight was adjusted to incorporate self-reported intention to vote. Respondents with a validated record of voting early in the 2024 general election received full weight. For the remaining respondents, four-fifths of the final probability that a registrant would vote in the 2024 election was based on the registrant's ex ante modeled turnout score, and one-fifth was based on self-reported intentions, based on prior Times/Siena polls, including a penalty to account for the tendency of survey respondents to turn out at higher rates than nonrespondents. The final likely electorate weight was equal to the modeled electorate rake weight, multiplied by the final turnout probability and divided by the ex ante modeled turnout probability.
Finally, the sample of respondents who completed all questions in the survey was weighted identically as well as to the result for the general election horse-race question (including leaners) on the full sample.
Weighting (unregistered Wisconsin voters)
The sample of unregistered voters in Wisconsin was weighted by The Times using the R survey package in multiple steps.
First, respondents were given a base probability of being unregistered, based on their self-reported registration status and whether they reported previously voting in Wisconsin at their current address.
Second, the sample was weighted to match Census-based targets for the demographic characteristics of unregistered voters in Wisconsin by age, race, education, sex and region.
Third, the probability that an unregistered voter in Wisconsin would vote was based on their self-reported intention to vote.
Finally, the unregistered respondents were weighted to represent 4 percent of the likely electorate, based on an analysis of L2 voter files with vote history and registration date after the 2016 and 2020 elections.
The margin of error accounts for the survey's design effect, a measure of the loss of statistical power due to survey design and weighting.
The design effect for the full sample is 1.27 for the likely electorate in Arizona, 1.60 for the likely electorate in Georgia, 1.37 for the likely electorate in North Carolina, 1.37 for the likely electorate in Nevada, 1.45 for the likely electorate in Michigan, 1.34 for the likely electorate in Pennsylvania, and 1.61 for the likely electorate in Wisconsin.
Among registered voters, the margin of sampling error is plus or minus 3.4 points in Arizona, including a design effect of 1.24; plus or minus 3.5 points in Georgia, including a design effect of 1.3; plus or minus 3.6 points in North Carolina, including a design effect of 1.37; plus or minus 3.4 points in Nevada, including a design effect of 1.25 ;plus or minus 3.5 points in Michigan, including a design effect of 1.29; plus or minus 2.8 points in Pennsylvania, including a design effect of 1.26; and plus or minus 3.6 points in Wisconsin, including a design effect of 1.72.
For the sample of completed interviews, among the likely electorate, the margin of sampling error is plus or minus 3.6 points in Arizona, including a design effect of 1.28; plus or minus 4.1 points in Georgia, including a design effect of 1.61; plus or minus 3.7 points in North Carolina, including a design effect of 1.37; plus or minus 3.8 points in Nevada, including a design effect of 1.42; plus or minus 3.9 points in Michigan, including a design effect of 1.45; plus or minus 3 points in Pennsylvania, including a design effect of 1.37; and plus or minus 3.7 points in Wisconsin, including a design effect of 1.36.
Historically, The Times/Siena Poll's error at the 95th percentile has been plus or minus 5.1 percentage points in surveys taken over the final three weeks before an election. Real-world error includes sources of error beyond sampling error, such as nonresponse bias, coverage error, late shifts among undecided voters and error in estimating the composition of the electorate.