1. Enormous distances: the polls draw different elections
Disagreement goes much further than a few decimals. The CIS of October places the PSOE 11.8 points ahead of the PP; NC Report, in its poll published on 9 November, places the conservatives seven points above the socialists. Between those two advantages there is a margin of 18.8 points.
SW Demoscopia, which broadcasts on the new channel La Séptima with a daily tracking, is the other institute in this comparison that gives first place to the PSOE: its report of 7 November attributes a 2.2-point advantage. In the rest of the selected studies the PP wins, although with distances that range from seven tenths to seven points.
Although dates matter, since the CIS interviewed in early October and the most recent measurements reach November, the political contrast remains: some results make it possible to talk about an open race and others show a wide advantage for Feijóo.
2. Maximums and minimums on a roller coaster: how much distance does each poll give
To give an example from nine studies brought together, the PP moves between 22.2% of the CIS and 32.6% of NC Report: 10.4 points. The PSOE goes from 25.2% of Sigma Dos to 34% of the CIS: 8.8 points. If the October reference from the CIS is removed, the distance between the extremes remains 4.3 points for the PP and 5.6 for the PSOE.
The decisive column for visualizing the disagreement is the advantage between them. GAD3 and 40dB. give the PP exactly the same percentage, 31.9%, but the first places it six points ahead of the PSOE and the second, only two.
On mobile, swipe the table to view all columns.
| Institute / outlet | Publication / report | Fieldwork | PP | PSOE | Advantage in points |
|---|---|---|---|---|---|
| NC Report / La Razón | Nov. 9 | Nov. 2–5 | 32,6% | 25,6% | PP +7,0 |
| GAD3 / ABC | Nov. 9 | Nov. 3–5 | 31,9% | 25,9% | PP +6,0 |
| Sigma Dos / El Mundo | Nov. 8 | Oct. 30–Nov. 6 | 31,0% | 25,2% | PP +5,8 |
| SocioMétrica / El Español | Nov. 8 | Accumulated since Sept. 1 | 32,0% | 27,7% | PP +4,3 |
| Ateneo del Dato / elDiario.es | Nov. 8 | Nov. 2–5 | 31,5% | 28,6% | PP +2,9 |
| 40dB. / EL PAÍS–SER | Nov. 9 | Nov. 2–4 | 31,9% | 29,9% | PP +2,0 |
| More in Common | Nov. 3 | Oct. 21–25 | 28,3% | 27,6% | PP +0,7 |
| SW Demoscopia / La Séptima | Report of Nov. 7 | Oct. 26–Nov. 6 | 28,6%* | 30,8% | PSOE +2,2 |
| CIS · reference of October | Oct. 14 | Oct. 1–6 | 22,2% | 34,0% | PSOE +11,8 |
Vote intention estimation percentages. The advantage is the difference between PP and PSOE, in percentage points. Selection of studies with different dates: it is not a time series nor an average.
* SW Demoscopia separates PP–UPN, with an additional 0,2%. The CIS, highlighted in yellow, corresponds to October. Sources: studies
*Differences in percentage points, calculated based on the published data. They do not represent voter movements between one poll and another nor a range from a single study. SW presents PP–UPN separately, with an additional 0,2%; the CIS corresponds to October.*
3. The outrage already has names: Tezanos, 'TelePedro' and the conservative press
The accusations have come to light. The socialist Rafael Simancas, secretary of State for Relations with the Courts, called on November 3 the “deception” of the seat projections of “the right-wing newspapers” with samples of 3,000 interviews or fewer, as reported by laSexta.
José Miguel Contreras, who promoted La Séptima, stated to verTele that the results he handled, from SW Demoscopia, had little to do with those disseminated by the “conservative press”. In the opposite direction, El Español has questioned SW’s credibility citing sources from the sector and underscoring the socialist track record of its head. Let us remember that several outlets refer to Contreras’s channel as 'TelePedro', for its affinity with the PSOE.
The discussion has moved from what the polls say to who makes them and for whom. All parties and their grassroots think that the studies from the media that are least affiliated are meant to 'depress' their electorate and that it is their own that are correct. In short: polarization has arrived at the polls.
4. Cook, cook... and the effect of jumping on the winner’s bandwagon
The so-called cooking kitchen turns answers into an estimate: it adjusts the sample, handles the undecided, and calculates what participation is likely. It is necessary work, but its decisions can change the result a great deal. The question is whether those adjustments are justified and explained.
Political temptation has a name: the *bandwagon* effect, or bandwagon effect. It consists in a candidacy gaining appeal because it appears as the winner or on the rise. Various experiments and political science itself explain that news about jumps in the polls could increase the declared support for the benefited parties.
That is why a survey can be used to try to create expectations of victory. There is no need to alter interviews to exploit that effect: one can also select the favorable data, amplify it, and present an uncertain evolution as a consolidated trend.
This does not mean that all cooking is fraud, nor that bandwagon effects always work. The charge of manipulation requires checking methodological decisions and results; a difference from other firms is not enough to prove it.
5. How to orient yourself: follow DEMÓCRATA’s average, don’t get stuck with a headline
A survey can change the day’s conversation. To know whether it is changing the trend, it is useful to observe whether the movement repeats across multiple studies and for how long. That is the usefulness of Demócrata’s aggregator:
](https://fren.democrata.es/embed/widgets/a3n-29n/widget/encuestas.html).
The tool brings together polls and calculates a daily average, giving more weight to the recent ones and to larger samples, with a single contribution per firm. It makes it possible to follow the evolution and check whether an upswing has continuity or remains isolated.
6. The backdrop of 23J: no one wants to confuse expectations and result
The memory of 2023 explains part of the tension. The PP won the elections, but its victory did not produce the majority that several polls were anticipating. The latest GAD3 report gave 180 seats to PP and Vox; its final result was 170. The 'shock' for the PP at not being able to access the Government, as the polls were advancing or the 'surprise' of the PSOE—to whom, whatever they say, the result caught them unprepared—hangs in the air today as never before.
There were warnings before voting: 40dB projected 173, below the absolute majority. That is why it is not accurate to claim that all the polls were predicting the same thing or that they failed in the same way.
The precedent leaves parties facing different risks. Feijóo combats excessive confidence: on October 8 he asked his supporters not to give any election as already won and reminded them that polls do not replace votes. Sánchez combats resignation: in Valencia he called on his followers not to give the battle up for lost.
The political reading is clear. It would harm the PP if its sympathizers considered voting unnecessary because victory seems certain. For the PSOE, that its own would stay at home because they believe it is impossible to make up the ground. 23J serves both as a warning, even though each one draws a different lesson from it.
7. There may be a simpler explanation: samples, dates, and seats
Not every divergence needs an underlying political operation. One phone survey, another online, and a measurement accumulated over weeks can capture different electorates.
Then comes the conversion into seats. Spain elects the Congress in 52 constituencies, not in a single national pool. A point of votes spread across many provinces can yield different results from the same support concentrated where the last deputy is contested.
The size of the sample matters, but it does not resolve everything. More interviews do not automatically correct a bias. Nor does Simancas’ critique, by itself, prove that a projection is invalid: you have to examine how the company combines its interviews, its territorial data, and its model.
8. The paradox that all DO conclude the same thing: the right will govern
The SW Demoscopia report contains a conclusion that may get lost behind the headline about the socialist victory. Its central breakdown attributes 128 seats to the PSOE, 116 to the PP, and 61 to Vox. The last two add up to 177; with the two from PP–UPN, counted separately, they reach 179.
That is, the same study allows us to state that the PSOE would come first and that PP and Vox could gather an absolute majority. These are two different questions: who gets more votes or seats, and who can form a Government.
The margin, for its part, changes a lot: GAD3 publishes 199 deputies for PP and Vox, and SocioMétrica, 202 including PP–UPN. A tight majority and a comfortable one do not offer the same negotiating power or the same capacity to absorb surprises. But the consensus is very broad: PP and Vox will break the absolute.
9. Polls move into the campaign: each party chooses the data it needs
The electoral use is already visible. After the premiere of La Séptima's poll, the PSOE circulated its results with the message "There's a contest on this 29-N".
Javier Ruiz had announced at the chain's presentation: "We intend to change the electoral campaign". The daily publication of estimates turns every variation into material for talk shows, social networks, and talking points. The PP, for its part, puts forward more subtly but clearly the same argument again and again: in the average of polls, Feijóo has been leading consistently for years.
Two headlines that seem opposed can also coexist. The 40dB poll, published this Monday, makes it possible to talk about a recovery for the PSOE and about a majority of PP and Vox. Emphasizing only the first conveys a comeback; emphasizing only the second conveys continuity of the right's advantage.
That's where another battle is fought: it is not always the number that's discussed, but rather which part is highlighted and what expectation is built around it. The risk for the reader is ending up choosing the poll that confirms their preference.
10. And what do you think? Make your forecast in Demócrata's simulator
Demócrata allows you to move from the polls to your own scenario. In the electoral simulator you can distribute the 350 seats in Congress and calculate which parties would reach the 176 of the absolute majority.
You can try a narrow majority, a result without sufficient support, or a change in the first position. The tool allows you to save and share the forecast and modify it until the polling stations close.
And in addition, you can share it on social networks, as many users already do. After all, and given the spread of the polls, with a little judgment, it’s a good time to do knockout predictions and beat all the -very expensive- opinion-polling institutes. Who will get the result of 29N right?