Where Do Pollsters Find 1,000 People?

Every political poll quotes '1,000 respondents.' But who are those people? Ask yourself: where would a pollster even start looking for them?

• A headline poll is built from about 1,000 people
• Those people don't appear out of nowhere
• We need to trace where they actually come from

Before we reveal the sampling frames, commit to your best guess. If a national pollster wants 1,000 U.S. adults, what is the most common way they start?

• Pick one initial hypothesis
• Your answer will be tested against real polling practice

The visual below summarizes the main recruitment channels real national pollsters use, based on public method statements. In the past decade, online opt-in panels and text-to-web have grown dramatically.

• Most modern national polls do not only cold-call random phone numbers
• Pollsters combine phone, online, text, and mail invitations
• Many respondents are regular panel members, not people found at random

The people a pollster tries to reach make up a sampling frame. A frame can be a list of cell numbers, addresses, voter files, or an online panel. Pollsters don't need to reach a perfect miniature of America; they collect about 1,000 voices and then weight the results so the respondents' ages, races, regions, and education look like the country.

• A sampling frame is the list or pool pollsters draw from
• Random-digit dialing and address-based sampling aim for probability selection
• Online panels and text-to-web are opt-in; weighting fixes imbalances

This explains national public polls. But boundaries exist: some pollsters use only probability samples, while others use pure opt-in panels; state polls or primary polls may use smaller samples and different frames. The recruitment strategy depends on budget, timeline, and the population being studied.

• National public polls often mix methods
• Some polls use probability sampling; others are fully opt-in
• A poll is only as representative as its frame and weights

Now apply the idea. Use the sliders to see what happens to a simulated poll when response rates drop or an opt-in panel skews younger.

• Change response rate and sample composition
• Watch the poll estimate move away from the true population value
• Ask: could weighting still rescue the number?

So where do the 1,000 come from? They come from a mix of frames: cell phone numbers randomly dialed, addresses mailed invitations, online panels where people signed up, voter files, and text-to-web links. Then the pollster weights the 1,000 to mirror the country. That's how a few conversations become a national headline.

• Recruitment starts from a frame: phone, address, voter file, or online panel
• About 1,000 respondents are never perfectly random in practice
• Weighting is the final step that makes the sample represent millions