Cluster17 · Spain
Cluster17 has published 2 polls in Spain this cycle, the latest on 25 August 2026. At 1 past election its final polls missed each main party (5% or more of the vote) by 1.5 points on average; it ranks 9th of the 53 pollsters we rate here (65/100).
Today
2 polls this cycle, the latest to 25 August 2026.
The score
- Accuracy: 33 of 50, from 1 final poll(s).
- The gap between the top two: 11 of 20.
- Openness: 21 of 30.
Accuracy and the gap are ranked against every final poll in this country; with few elections on record they are pulled towards a typical one. How the score works
Each of its polls minus the average of the other pollsters' latest polls at the time, party by party. Above zero, it had the party higher than the rest of the field. A lean is not an error: nobody knows who is right until polling day.
The gap between PP and PSOE
At past elections its final polls put these too high (+) or too low (−) on average: PSOE −2.8, Vox +1.3, Sumar +1.2, PP +0.5.
| Election | Average miss | Gap miss | Biggest misses |
|---|---|---|---|
| 2023 14 Jul 2023 | 1.45 | +3.3 | PSOE −2.8, Vox +1.3, Sumar +1.2 |
How it polls
| Method | online (CAWI) |
|---|---|
| Sample | typically 1,847 (from 1,637 to 1,847) |
| How often | about 0.2 a month |
| For | Agenda Pública |
French firm; Spanish polls for Agenda Pública and Le Grand Continent. Its published vote shares are a weighted direct intention (weighted by stated likelihood to vote) "sin estimación ni redistribución de indecisos" (Agenda Pública, Aug 2026).
What it publishes
| Raw or unprojected figures Shows what respondents said before the pollster's own adjustments · Y |
Yes | 8/8 |
|---|---|---|
| Full tables published Sub-groups and weighting can be checked · Agenda Pública ficha: CAWI, n, dates, weighting (incl. vote recall 2023 and 2024), margin; published shares are a turnout-weighted direct intention without redistribution of undecided voters. |
Partly | 3/8 |
| Method note How the sample was drawn and weighted · Y |
Yes | 4/4 |
| Sample size The basics for judging a poll · Y |
Yes | 3/3 |
| Undecided share Shows how firm the figures are · N |
No | 0/3 |
| Fieldwork dates When the poll was taken · Y |
Yes | 2/2 |
| Member of an industry body Signs up to its standards |
Not reviewed | 0/2 |
What its own adjustments do
It doesn't publish figures before its own adjustments, so we can't show what they do.
Figures as of . Published by Polling Forecast.





