Midterms 2026: Five Senate Races Where the Civic Data Tells a Different Story
Polling averages say one thing. Town-hall attendance, constituent-response rates and FOIA backlogs say another. We ran the numbers on five toss-ups.
Polls measure what voters say. Civic data measures what officials do. Ahead of November, we compared five toss-up Senate races on three behavioural metrics drawn from public records and Nject activity: constituent-response rate, town halls held per quarter, and median delay on public-records requests routed through their offices.
What the responsiveness gap predicts
In the 2024 cycle, incumbents in the top quartile for constituent responsiveness outperformed their polling average by 1.8 points on election day. The effect was strongest in races decided by under 3 points — exactly the kind of races on this list.
The pattern to watch
Two of the five incumbents have stopped holding open town halls entirely, substituting tele-town-halls with screened questions. Historically that substitution shows up six months later as a soft drop in turnout among their own base. Challengers who fill the vacuum with in-person events — both are doing so — convert at measurable rates.
Response rate range across the five offices: 31% to 84%
Median FOIA-style request delay: 12 to 61 days
Open town halls last quarter: 0 to 7
We will rerun this analysis monthly until the election. The full dataset is downloadable in the Data & Transparency section.