African American Research Collaborative


The 2022 Midterm Voter Election Poll, led by the African American Research Collaborative (AARC) in partnership with leading pollsters, is a large-scale poll of Black, Latino, Asian American, Native American and White voters nationally as well as in key states and congressional districts. Overall, the election poll targets 12,208 voters in the final days before the November 2022 election, offering exclusive information about the electorate including Congressional, Senate and Gubernatorial vote choices, issue priorities, evaluations of both major parties and what issues motivated voter turnout. The overall poll has a margin of error of 0.9%.

The poll is implemented for all racial/ethnic groups nationwide and also includes robust samples of African American and Latino voters within specific battleground states to provide reliable vote choice estimates in 11 states and nationally. States included are: Arizona, California, Colorado, Florida, Georgia, Michigan, Nevada, North Carolina, Pennsylvania, Texas, Wisconsin.

Respondents are selected from the voter file, and confirmed to be registered to vote. They are randomly contacted for an interview by telephone live-caller, SMS text message, email or online panel. The survey is available in English or Spanish at the discretion of the respondent. The poll was in the field during early voting for both vote-by-mail and early in-person voting, and continued through November 7th. Voters were confirmed to have already voted early, or in the final days, were 100% certain they were voting on Election Day. Weights are employed to balance the sample for demographic characteristics, with allowance for differential turnout rates, based on historic turnout data and current turnout models.

Margins of error:


Overall national profile +/- 0.9%

African American +/- 1.4%

Latina/o +/- 1.3%

White, non-Hispanic +/- 2.8%

Asian American +/- 4.0%

American Indian +/- 4.4%

Our voter poll uses an approach now also embraced by the national media exit polls of starting interviews with early voters before the election, and then also sampling election day voters. Our team pioneered this methodology in 2010 and demonstrated its accuracy through post-election voter file validation. This allows a statewide randomized representative sample and avoids the errors of cluster-sampling specific precincts in election day precinct exit polls. Our poll still offers a considerable advantage over national exit polls in that each racial group is weighted to match demographics for their specific group within each state. The national exit polls only provide state level, or national weights, and do not apply weights to racial groups. This has led to a socioeconomic bias in Black and Latino samples.