Not every community question is an election. Most aren’t. A city proposes a rezoning. A transit board floats a fare increase. A district considers closing a branch library. Each of these produces a room — a town hall, a comment period, a public meeting — and rooms are notoriously bad samples of the community they claim to represent.
The people who show up are the intense minority. The synthetic version lets you hear the rest.
The setup
Take a real decision with a defined affected population. Synthesize that population — not the state, not the county, but the specific catchment — and put the proposal to it as you would in the room:
- A plain-language description of the proposal.
- The support / oppose / unsure question.
- The open-ended “what’s your biggest concern?”
- A second pass after showing the strongest counter-argument.
That fourth step matters. Public meetings are persuasion events, not just measurements. Simulating the before-and-after tells you which objections are load -bearing.
An illustrative run
We modeled a hypothetical fare increase for a mid-size transit agency — a $2.75 → $3.25 base fare — against a synthesized ridership-weighted population.
| Group | Support (before) | Support (after argument) |
|---|---|---|
| Daily commuters | 31% | 44% |
| Occasional riders | 52% | 58% |
| Low-income pass holders | 12% | 19% |
The counter-argument shown was a reduced-fare program for pass holders. It moved every group — but the pass-holder group, the one the loud room would be angriest on behalf of, moved the least in absolute support while moving most in relief of stated concern. The objection wasn’t the price. It was trust that the discount would materialize.
“Fine, but I’ve heard ‘reduced fare program’ before. Show me the enrollment form.” — synthetic pass holder
A real town hall might never surface that cleanly, because the person who’d say it is working a second shift and not in the room.
The point
You do not run this to replace the meeting. You run it to walk in knowing which concerns are real, which are proxies for something else, and which argument actually moves the population you’re accountable to — not just the population that showed up.
This run is illustrative and uses a synthetic population; figures are not drawn from any real agency’s ridership.