About & Methodology

The 2025 study will feature a national sample stratified by state, conducted online with a single 20-minute questionnaire (10 minutes common content, 10 minutes team content). Each participating team receives a 1,000-person survey covering their custom content plus access to the full Common Content dataset. Team data remains embargoed for participating teams' private use. The cost is $14,000 for a 1,000-person sample, with additional cases available at $11 each. Key dates: team questionnaires due August 2025, survey fielded November 2025, data delivered March 2026. To participate, contact Marissa Shih (marissa.shih@yougov.com) at YouGov.

The Cooperative Election Study (CES) is conducted by YouGov using a matched random sample methodology. In even years, the survey is fielded in two waves: pre-election (typically October through early November) and post-election (November through December). In odd years, the survey is one wave that occurs in November. 

We employ YouGov's matched random sample approach, which draws a probability sample from the target population (using American Community Survey data) and then matches each target respondent with the most similar available respondent from YouGov's opt-in panel. Matching is conducted using a weighted Euclidean distance metric based on registration status, age, race, gender, and education. This produces samples that mimic the characteristics of random probability samples while being more cost-effective than traditional sampling methods.

The sample is weighted using entropy balancing to match American Community Survey distributions on key demographics (gender, age, race, Hispanic origin, education) and their interactions. Weights are then post-stratified by additional variables including voter registration status, vote choice, and born-again status. Final weights are trimmed and normalized to equal sample size.

Quality controls are applied to completed interviews before matching to the target frame. The large sample sizes enable analysis of legislative constituencies and rare events while maintaining reasonable precision across most states.

Common Content consists of approximately 15 minutes of questions, with about 10 minutes in the pre-election wave and about 5 minutes in the post-election wave. These questions are included on all administered surveys and amount to a 60,000 national adult sample survey. Common Content is asked at the beginning of each survey.

In addition to these questions, YouGov provides demographic indicators, party identification, ideology, and validated vote.

Team Content arises from institutions who purchase modules to the CES. Each research team that wishes to be involved in the project purchases a 1,000 person sample survey which is connected to the Common Content. Common content is asked of everyone, and then each individual team determines the other half of the questions asked of its 1,000 person sample. This amounts to 10 minutes of team content on the pre-election questionnaire and 5 minutes on the post-election questionnaire. A separate dataset is produced for each team’s content.

Vote validation

The CES matches respondents to voter files for purposes of validation of survey responses, including registration status, vote history, and party registration. CES offers validated registration and vote history for every election-year survey conducted since 2006.