U.S. desk · Proprietary instrument
U.S. Latino Radar
The races where Latino electoral weight, competitiveness and control of Congress intersect.
The index combines seven signals without reallocating weights. It is not a battleground-only list: a safe seat can matter through its electorate, representation or role in control of Congress.
Updated · 2026-09-16 · 2020-2024 ACS 5-year
races · average index 18.6
races · average index 30.4
Preliminary signal · no Vote-Scope projection
National index
Latino Vote Index
The two-party Democratic share of the Latino House vote, aggregated from all-Latino polls and published with its own precision.
95% confidence interval: 62.2% – 68.0% Two-party margin D+30.2
Three usable waves, 2 houses, and the most recent fieldwork is 125 days old. The effective sample behind the number is 1025 respondents once the recency discount is applied. This is an average in the arithmetic sense, not yet in the statistical one.
The two houses in the series sit 5.2 points apart on the headline metric. With this few houses, the LEVEL of the index partly reflects who fielded most recently: do not read a change of house as a movement of opinion.
We draw no trend line: three points from two houses would describe the gap between the houses, not the electorate.
Houses handle undecideds differently: in May 2026 UnidosUS and Equis both put the Democrat at 54 but the Republican at 27 and 35. The two-party share D/(D+R) is the only form comparable across houses.
Three tracks, never mixed
U.S. pollsters field at least two populations — adults and registered voters — and switch from a generic to a named ballot as the election nears. Both are valid; they are not the same measurement. Measured on our own data: 66.1% on an adults/generic wave ending 31 August, 56.3% on a registered-voter/named wave ending 8 September, same house, eight days apart. The index therefore keeps separate tracks and never averages across them.
| Track | Population | Ballot | Two-party | Waves | Field | Movement within track |
|---|---|---|---|---|---|---|
| Dedicated Latino polls Headline track | Registered voters | Generic | 65.1% | 3 | 125 d ago | +1.7 over 204 d |
| General-poll crosstabs | Registered voters | Named | 60.7% | 2 | 2 d ago | Fewer than three waves |
| General-poll crosstabs | Adults | Generic | 62.7% | 6 | 16 d ago | +2.1 over 217 d |
The waves behind the index
Each wave is compared against the national generic-ballot average over the SAME field window, never against the average on publication day.
| House | Field | Sample | Published D / R | Two-party | Gap |
|---|---|---|---|---|---|
| Equis Research | May 11, 2026 – May 25, 2026 | 2,000 ±2.0 | Margin only — kept out of the share index | — | — |
| BSP Research & Shaw & Company Research UnidosUS | Apr 27, 2026 – May 14, 2026 | 3,000 ±1.8 | 54 / 27 | 66.7% | +14.5 |
| Equis Research | Mar 10, 2026 – Mar 20, 2026 | 2,000 ±2.0 | 54 / 35 | 60.7% | +8.7 |
| Equis Research | Oct 15, 2025 – Oct 29, 2025 | 2,000 | No ballot published | — | — |
| BSP Research & Shaw & Company Research UnidosUS | Oct 8, 2025 – Oct 22, 2025 | 3,000 | 52 / 28 | 65.0% | +13.6 |
Race
214 eligible races. Rankings recalculate after every new House or Senate model run.
30 / 214
Methodology
Each component is scored from 0 to 100 and multiplied by its published weight. Incomplete races are excluded, never reweighted.