Party colors show sourced affiliation; gray means other or unmatched. Missing polls never mean zero support.
Hourly refresh scheduled. Feed checks through . Source details ↓
POLLING · SURVEYS, NOT PREDICTIONS
National congressional preference
546 survey records
Individual surveys of 2026 contests, including earlier cycle history from 2025. Matchups and populations stay separate; older or hypothetical matchups do not establish who is on the current ballot.
Simple average, not a forecast
Each pollster’s latest general-election survey ending within the last 30 days, with equal weight. Reviewed candidate identities join spelling variants; different matchups and voter populations stay separate. Percentages describe reported support, not a chance of winning.
Newest fieldwork appears first. When dates tie, the default favors more pollsters not marked internal or partisan in this 30-day window. Source labels remain visible; this is not a pollster-quality rating.
Registered voters · 1 poll
Sep 30, 2026–Oct 4, 2026 · Lead between the two highest answers: 8.0 percentage points
Add each answer’s reported support across these 1 surveys and divide by 1. No house effects, sample weights, undecided allocation or error band is applied. Incomplete numeric matchups are excluded from this calculation but remain in the survey history. Full methodology →
Likely voters · 3 polls
Sep 16, 2026–Oct 2, 2026 · Lead between the two highest answers: 7.0 percentage points
Add each answer’s reported support across these 3 surveys and divide by 3. No house effects, sample weights, undecided allocation or error band is applied. Incomplete numeric matchups are excluded from this calculation but remain in the survey history. Full methodology →
Chart and reports: observations 401–450 of 546. Averages above use all eligible surveys, independently of this history window. Use the page links to inspect every earlier observation.
No recent polls available. The history below ends Oct 20, 2025; “recent” means fieldwork ended within 30 days.
Hover or tap a date to see the survey and source.
DemRep
Party colors: Republican red, Democratic blue, Independent purple. Other or unmatched affiliations are gray. Dash patterns distinguish candidates sharing a color.
Dots show reported support. Lines connect surveys over time—not an average or forecast. Pollsters can differ. Missing values and multiple surveys ending on the same date break the lines.
How to read this chart
No average, trend estimate, or win probability. Different matchups and populations are not combined. Dates mark the last day people were surveyed. The vertical scale always starts at zero; use “Full 0–100% scale” to compare charts on the same scale. The number of people surveyed is the survey sample size, not the number supporting a candidate. Select “Survey source” to inspect every report, including overlapping dots.
Latest surveys · selected matchup
Individual reports. Original sources and methodology below.
Labels and figures are supplied by VoteHub. A listed matchup does not establish ballot qualification or primary/general-election status unless the source explicitly says so. Unlisted responses may include undecided voters or other choices; the displayed figures are not rescaled to total 100%.
Polling compilation: VoteHub · CC BY 4.0. Electionpedia normalizes dates, matches races, and visualizes reported results; it does not estimate outcomes. Original survey links are attached to each report.
Request-time API fallback; scheduled storage is unavailable. Across all feeds, the latest survey ended 2026-10-20. Check the field dates above for this selection. Feed update status →
Last successful feed updates
Senate: 10/21/2026, 10:30:48 PM UTC · 491 survey records
House: 10/21/2026, 10:30:48 PM UTC · 154 survey records
Governor: 10/21/2026, 10:30:48 PM UTC · 371 survey records
National preference: 10/21/2026, 11:06:42 PM UTC · 546 survey records
An update time is when we checked the feed, not when a survey was conducted. Coverage is not exhaustive.
Generic-ballot surveys ask about party preference nationally. They are not polls of any individual district or candidate.
PREDICTION MARKETS · NOT POLLS
What are markets pricing in?
Polymarket widgets
These percentages reflect trading in event contracts—not the share of voters supporting a candidate or an Electionpedia forecast.
These contracts concern national chamber control, not an individual race. Separate contract prices need not add to 100%. A “No” price is not automatically the opposing party’s price.
Prices are supplied and updated directly by Polymarket. If a widget is blocked or unavailable, use its source link; Electionpedia does not substitute a cached price or zero. Trading buttons are disabled. Widgets load automatically near the screen and contact Polymarket; privacy details.
Sources & coverage
How to read these polls and what’s covered
House and Senate race pages come from the existing 2026 FEC filing directory, not certified ballot lists. Governor pages cover all 39 scheduled state and territory contests. Candidate ballot qualification is established only by the linked election authority, independently of polling.
Recent means fieldwork ended in the last 30 days. Latest polls sorts by fieldwork end. Closest races sorts by the margin in a recent simple average, using each pollster’s latest general-election survey; missing averages go last. Matchups and surveyed populations stay separate. Each race page shows the included sources and the average rule.
VoteHub’s API does not currently include collection mode or reported margin of error. Those fields say “Not supplied in this feed,” with a link to the original report. We do not derive uncertainty from sample size alone. See AAPOR’s disclosure guidance ↗.
A scheduled job checks the structured VoteHub API hourly, validates each feed, and saves the last good data independently of page visits. Failed or unexpectedly truncated feeds do not replace those records. Coverage is not exhaustive; missing polls never mean zero support. Check feed health →
Electionpedia.com
