Congress Cut Its Town Halls by 63 Percent. Then It Reversed Course. The Format Was Never the Problem.

Members of Congress held 3,480 town halls in 2017. By 2024 they held 1,023, the lowest total in LegiStorm's database. Then 2025 reversed it — 1,772 events by late September, 993 of them in person, the most since 2019.

Which brings us to August, the month the format was built for. Members are home, districts are accessible, and the calendar clears for five weeks. As of August 10, only eleven members had a town hall scheduled for the rest of it.

Eleven, out of 535.

Even if more announce late, that leaves 524 members with nothing at all on the public calendar.

None of that is a decline. It's an oscillation, and a listening mechanism that swings this hard inside two years isn't really a mechanism.

Nobody Stopped Holding Town Halls Because They Stopped Caring.

In a study of 10,918 town halls across the 115th through 117th Congresses, roughly 78 percent of House members held any at all — meaning better than one in five held none. And it isn't one decision but two: whether to hold them, then how often and in what form. Both get made under risk.

Every communications director knows that calculation without having to run it. One hostile exchange, filmed on a phone held over someone's head, outlives every substantive answer the member gave that night. The event lasts sixty minutes. The clip does not. Weighed that way, moving August to district visits and roundtables isn't timidity. It's a rational read of an asymmetric bet.

The problem is that the other side of the bet gets priced at zero, and it shouldn't be. An office that holds no open events stops hearing anything it didn't already know. It doesn't feel like a cost, because nothing happens. That's exactly what makes it expensive.

Anything optional eventually gets skipped. The offices that keep hearing from constituents through a difficult year are the ones where listening is wired into a program that runs on a calendar, rather than into an event that requires a fresh act of nerve every quarter.

The Mass Telephone Town Hall Solved the Wrong Problem.

Offices have already moved, and the numbers show it. Of the 1,772 events held through late September, 993 were in person — meaning roughly 780 were not.

The mass telephone town hall is the workhorse in that second group. Dial tens of thousands of constituents, hold whatever fraction stays on the line, screen the incoming questions, let the member talk for an hour.

It solves the room. It does not solve the listening. The audience is as undifferentiated as a mass email — everyone dialed, no one chosen — so whatever the member hears is a function of who happened to be home and stayed on. The researchers who have studied the format most closely note that it draws persistent criticism for failing to facilitate engagement and accountability, and that critique is fair to the version most offices run.

It also doesn't retire the risk the way offices hope. The calls get recorded. Hostile questions clear the screen. And the screening itself generates a second story, because a constituent who didn't get called on has a phone too, and "my office ducked me" travels about as well as the exchange would have.

So the office spends the budget, keeps most of the exposure, and finishes the hour knowing roughly what it knew going in.

The People Who Show Up Are Not the People You Represent.

Dr. Michael Neblo testified before the Committee on House Administration in December, drawing on a survey of 1,001 adults fielded last fall. Roughly one in seven constituents actually contacts a member in a given year. Among those who don't, more than 40 percent say they believe members do not care what they think, and more than 47 percent say contacting an office is simply too difficult or time-consuming. Pew found in 2024 that more than 80 percent of Americansbelieve elected officials don't care what people like them think.

So the room at an open town hall is filled by the narrow slice of constituents motivated enough to overcome all of that. That group is real, and it deserves to be heard. It is also not the district.

The finding that should interest a communications director is a different one. When constituents are recruited into structured forums, previously disengaged citizens prepare, show up, and contribute more than the people who already call the office. Ninety-four percent said the experience was very valuable and that they would do it again — in an era when roughly 13 percent say they trust Congress as an institution.

The problem was never the format. It's that the invitation list is whoever happens to see the notice, or on a mass dial, whoever happens to pick up. The most consequential decision about the event gets handed to chance.

You Cannot Eliminate the Hard Question. You Can Change What Kind You Get.

Concede the part most vendors skip. No format makes the hostile question go away, and an office that engineered a room where it couldn't happen would be holding an event not worth holding. Constituents can tell. So can anyone who later watches the clip.

What the research changes isn't whether the hard question comes. It's whether it's answerable.

Claire Abernathy, Kevin Esterling, William Minozzi, Michael Neblo and colleagues ran a field experiment in which four members of Congress joined their constituents for telephone town halls. Participating improved constituents' view of the format and of the member. More usefully, the researchers tested a specific reform — a single-topic focus with briefing materials distributed in advance — and it improved how constituents evaluated the event. The Congressional Management Foundation's guidance converges on the same variables: recruit toward a representative cross-section, use a real moderator, and send two to five pages of neutral background beforehand. When participants got that background, CMF found correct answers on a policy quiz rose by half.

That is the entire mechanism, and it is worth being concrete about what it buys. A pointed question from a constituent who read four pages on the reimbursement formula is a hard question with an answer. A question from someone who saw a headline that morning is an ambush — unanswerable in thirty seconds, not because the member is wrong, but because the premise has to be unwound before anything true can be said about it.

One of those is a bad two minutes. The other is close to the best thing that can happen at one of these events: a member answering a real challenge well, in front of people equipped to recognize that the answer was good.

None of which makes it a friendly room. Recruiting toward a representative cross-section is the recommended practice precisely because it corrects for self-selection toward the already-motivated. Curating for topic is not curating for agreement, and the offices that confuse the two get the worthless version of both.

Build the Program. Then Schedule the Event.

Know who cares about what before you need them. Invite the ones the topic is for. Run the event on a defined process instead of improvising it.

The dependency is the part most offices miss. You cannot build a topic-segmented invitation list in the two weeks before an event. Segments are a byproduct of a program that has run long enough to learn who responds to what — which is why the offices that can convene a useful virtual event on short notice are the same offices that were already sending differentiated constituent mail all year.

Who gets invited is a segmentation question. What they read first is a content question. Whether the right people show up is a deliverability question. On one of the Senate programs we manage, that groundwork is more than 15 audience segments and better than 12,000 survey responses collected through new-contact engagement sequences. An engaged room gets built the same way an engaged list does.

For the offices that engage their constituents through virtual town hall events with us, we run a nine-step process built to do two things: line up the right audience, and make sure the sixty minutes that follow produce something the office can actually use. Every decision it calls for gets made weeks out, not improvised the night of.

The Format Was Never the Point.

An office that knows which constituents are tracking a hospital closure and which ones have never written in at all can put several hundred of the right constituents in a virtual room on two weeks' notice, all of them there for the topic, and come away knowing something it didn't know before. The office without that groundwork is choosing between an open door, a mass dial, and an empty calendar. Most chose the calendar this year.

That only looks like the safe option. An office that holds no events doesn't avoid the hard question. It just hears it later, somewhere it can't be answered.

Fall calendars get set over the next few weeks, and the segmentation behind a room like that takes longer than that to build. If you want a read on what your office already has and what it would still need, send us a note.