The Center for Sociological Research (CIS) has published what may be its last preelectoral poll under the presidency of José Félix Tezanos. And once again, it remains the dissonant note of all demoscopic work, as it calculates that the PSOE would be the leading electoral force in the country, with 33.3% of the votes and a seat range of 134 to 143. Behind them, the Partido Popular (PP) would obtain 25.5% of the votes and between 98 and 106 seats. Vox, for its part, would gather 18.4% of the votes and 57 to 65 seats. Further back, the Broad Front (FA) would be between 7 and 11 seats, with an estimated vote share of 4.7%, to which one would have to add the support of other candidates within this political space, such as Compromís.
There is also the difference that the CIS represents compared to private polling firms: between November 8 and 9, numerous surveys were published by Sondaxe, IMOP, ElectoPanel, SocioMétrica, Sigma Dos, GAD3 or 40dB, among other companies. In all those that offered a seat estimate, the sum of PP and Vox exceeded the absolute majority, in some cases nearing 200 deputies. By contrast, the CIS leaves both parties with a maximum of 171 seats, even when adding the upper ends of their ranges. In this preelectoral poll, the public body conducted more than 45,000 interviews between October 7 and 30 — that is, with fieldwork that includes the Urda rail accident.
However, a poll of this size does not have to be more accurate than a smaller one. As some experts point out, a very large sample with a lot of bias can be worse than a smaller one with less bias.
The problem of vote recall
As is well known, the CIS’s estimates carry a series of biases that invite caution, and it is common to see additional calculations beyond those made by the center itself. Beyond the so‑called “kitchen,” the CIS has a problem with vote recall. How is this measured? By comparing what respondents say they voted for in the last elections with the results that actually occurred.
This time, among those who remember voting for a candidacy in 2023, 42.5% state that they did so for the PSOE, but the Socialists obtained 31.9% of the votes for candidacies. It is a difference of more than ten points. With the PP the opposite happens: it represents 24% of that recall, compared with 33.3% that it obtained at the polls. Vox also appears below its actual result: 10.6% versus 12.5.
“This time, among those who remember voting for a candidacy in 2023, 42.5% state that they did so for the PSOE, but the Socialists obtained 31.9% of the votes for candidacies”
There is another way to look at it. In the recall declared by the respondents, the PSOE exceeds the PP by more than eighteen points. Yet, on June 23, the PP won the elections by 1.4 points. While recall does not have to be exact, the declared composition is very different from what the ballots left. And that can lead to important consequences for vote estimation.
What happens if we correct that imbalance?
Although we do not yet have the individual microdata, we can perform a simple reestimation by adjusting vote recall to what actually happened in 2023. The procedure consists of giving each group of former voters a weight proportional to the result their party obtained in those elections, while preserving the preferences they declare today in the survey.
According to this model, the PSOE would obtain 29% of the votes, the PP 27% and Vox 21.7%. This would considerably reduce the gaps with respect to the CIS’s official November 2026 estimate and would leave the PP only two percentage points behind the Socialists. Together, PP and Vox would sum 48.7% of the votes, compared with the 43.9% attributed to them by the CIS. If we account for changes by party, the PSOE would lose 4.3 points relative to the official calculation, while the PP would gain 1.5 and Vox another 3.3.
“Together, PP and Vox would sum 48.7% of the votes, compared with the 43.9% attributed to them by the CIS”
But there is another question: how would that distribution translate into seats? To approximate, we applied the national changes to the provincial direct vote and then allocated the seats using the D’Hondt method. The result is 116 seats for the PSOE, 112 for the PP and 80 for Vox. To these two would have to be added the two seats obtained by the UPN-PP candidacy in Navarre. In total, 194 seats, eighteen above the absolute majority.
The importance of always using the same scale
When we want to see if our diet is working, it is important to always use the same scale because, even if it may not measure our weight precisely, it will reflect changes well. With the same logic, when we analyze polls, it is worth comparing the estimates of the same company to see what is happening with the vote. And here the CIS also offers striking data.
In October, the agency placed the PSOE at 34% of the votes, the PP at 22.2% and Vox at 17.1%. A month later, the PP has gained 3.3 points and Vox another 1.3, while the PSOE has fallen by seven tenths. The sum of PP and Vox thus rises from 39.3% to 43.9%, a gain of 4.6 points, while the distance between PSOE and PP narrows from 11.8 to 7.8 points.
“The sum of PP and Vox thus rises from 39.3% to 43.9%, a gain of 4.6 points, while the distance between PSOE and PP narrows from 11.8 to 7.8 points”
Even if we take the official estimates at face value, the CIS shows a simultaneous advance of PP and Vox. Until now, much of Vox’s growth had been interpreted in relation to the losses of the Popular Party (what we call intra‑bloc transfers). But here both parties rise at the same time.
To compare the left space between the two polls, it is necessary to add to the Broad Front the candidacies that previously formed part of Sum ar, such as Compromís, Ara Més and UPC, in addition to Podemos and Adelante Andalucía, parties that could be part of a PSOE majority. In this way, the whole would move from 43.2% in October to 42.8% in November, a drop of barely four tenths. The difference between both blocs would have shifted by five points in just one month: from a left-leaning lead of 3.9 points to a right-leaning lead of 1.1.
Where the votes come from
To better understand these movements, it is worth examining the transfers that the survey captures. Among those who remember voting for the PP in 2023, 63.2% declare they would vote for it again, but 20.1% say they would now vote for Vox. The opposite movement is much smaller: only 8% of Vox’s former voters would choose the PP, while 75.7% would continue supporting Santiago Abascal’s party.
The difference is bigger than it seems, because the PP obtained more than 8 million votes in the last general election, compared to just over 3 million for Vox. Therefore, that 20% of former PP voters represents a much larger pool of electors than the 8% who declare the opposite movement. One of the keys to Vox’s evolution lies precisely there—in its ability to attract part of the PP electorate without losing too much of its own support.
In the PSOE, declared loyalty sits at 66.1%, though 14.1% of those who voted for it in 2023 still do not know what they will do. That share of undecideds will be very important in the two weeks of campaigning.
“One of the keys to Vox’s evolution lies in its ability to attract part of the PP electorate without losing too much of its own support”
The situation is even more fragmented among the former voters of Sum ar: 21.6% say they would now vote for the PSOE, 21.4% would opt for the Broad Front and 20.6% remain undecided. In this sense, the space of the alternative left shares a significant difficulty in retaining the vote it won three years ago.
In recent days, spurred by the controversy around interviews with presidential candidates on Jordi Wild’s podcast, much has been said about young voters. In this poll, the CIS shows notable differences in this bloc. Among 18 to 24-year-olds, Vox is the leading force with 22.1% of direct intention, ahead of the PSOE with 19% and PP with 12.2%. Among those aged 25 to 34, a similar pattern repeats. By contrast, from 65 onward, the Socialists exceed 40% and Vox falls below 7%. We are not, therefore, talking about identical electorates, nor is it expected that they react in the same way to campaign movements.
“Among 18 to 24-year-olds, Vox is the leading force with 22.1% of direct intention, ahead of the PSOE with 19% and PP with 12.2%”
In 2023, although it continued to overestimate its vote, the CIS was the poll that best estimated the PSOE’s result. Now two weeks lie ahead where the voting trend could change, but behind the agency’s data there appears a majority for PP and Vox and a growth trend for these two parties in the last month. There is still a lot at stake and the scenarios are open, but the main conclusion left by this demoscopic work, CIS included, is that the board is tilted more to the right than the one we found three years ago.
