Draw a population.
Ask it anything.
We model any demographic slice you can define, then answer your questionnaire the way that slice would answer it. Weeks of fieldwork, returned before you finish reading the question.
Tap any block to change a variable, then run the wave.
Demo output. Figures show the report format, not a fielded poll.
Three steps, one wave.
Draw the slice
Gender, age band, ethnicity, education, household income, state, urbanicity, party ID, marital status — plus religious attendance, home ownership, union household, parental status and 2024 recalled vote.
Field the questionnaire
Paste the wording you would send to a field house. Forced choice, Likert scales, favourability, open-ended. Question order and framing are preserved, because they move the answer.
Read the crosstabs
Distributions, banner points, and a calibration band on every item. Change one variable and re-field — a whole wave costs seconds, so you can afford the follow-up you would normally cut.
Same deliverable. Different clock.
Calibrated against real polls, and reported with its error.
Each slice model is fit on public opinion data — ANES, the General Social Survey, Pew's American Trends Panel, the Cooperative Election Study — then held against benchmark waves it never saw in training. Marginals are weighted to Current Population Survey targets on gender, age, ethnicity, education and region, the same frame your field house uses.
Every answer ships with the calibration error measured for that specific slice, so you know which numbers to lean on and which to read as a signal.
- Not a replacement for your benchmark wave.Field one real poll a quarter. We calibrate to it.
- Not a read on breaking events.Anything after the data window needs live respondents.
- Not a person.It reports how a slice tends to answer. It does not know what anyone thinks.
Poll the slice you could never afford to reach.
Private beta. Campaigns, brands and newsrooms first.