Polling & Analytics
September 2026
Data Principles
Sorting through the ever-growing list of public opinion polls can be complicated and tricky. Widespread skepticism of data and polls is, frankly, entirely warranted by their frequent misuse. Despite that, Searchlight firmly believes that polling can still be a valuable way of listening to Americans and understanding their opinions. It’s one of the only tools we have to talk to ordinary people who are disengaged from the political process.
To make sure our work is useful, Searchlight is committed to a set of principles that hold us to a high standard. We aren’t here for business as usual, and our principles reflect that.
Shared Facts
- We believe that discussions about policy need to start from agreed-upon truths about the world. Polling is one way to get at truths about opinion and belief, and to supplement a broader universe of evidence.
- Our polling and data work seeks to provide useful public evidence as a foundation for discussion. Reasonable people can disagree on interpretations and conclusions, but high-quality foundational information makes everyone’s position better.
Learning
- We want our polls to provide new information, not tell us what we want to hear.
- There’s a lot we don’t know about the American population and electorate. Every poll is an opportunity to learn more. We think it’s a waste to run a poll that can’t surprise you, and our ultimate goal is to expand our understanding.
Humility
- Polling is hard, and we don’t expect to get it right every time.
- We accept that sometimes our beliefs will need to change when new evidence comes in. Sometimes a best-effort polling question ends up, on reflection, being misleading or just plain inaccurate. We know this is a part of the process, and we’ll be up front when it happens.
- We also accept that the public’s beliefs can change. We will return to questions and check our own assumptions over time to ensure our conclusions continue to be true.
Nuance
- We know opinion is complicated, and a single question isn’t the end of the discussion. Survey questions are short and can only tackle one aspect of an issue. (Maybe voters love your policy, but hate the idea of paying for it, for example.) We try to honor this complexity by tackling questions from multiple angles. Further, we believe that a single survey result will never be the end of a discussion, and that we should always be willing to learn more.
- We won’t cherry-pick questions, and we’ll always endeavor to see the broader context.
Transparency
- When running polls, we ask a lot of questions. It can be tempting to pick out the ones that came back “good” — i.e., in the direction you were hoping they would trend — and quietly bin the rest. To avoid this, we will always release our full questionnaires. (For omnibus-style polls, we’ll roll results out in chunks by topic and post the full questionnaire at the end).
Methodology
- While we do design our own surveys, we don’t field or weight our polling in-house. (Hey, we’re a small team!)
- We will always work with our polling partners to ensure we are accurately representing our respondent pool, whether that’s all American adults, the electorate, or a more targeted slice of the country. At a minimum, our crosstabs will include the number of interviews conducted, the modes used to contact respondents, and when the survey was conducted.