Vox Voters Regretted Their 2023 Vote the Most: They Voted Expecting a Government That Never Came

In June 2023, two out of three Spaniards took a PP win for granted, and so did nearly half of PSOE voters. After the 23J election, 13.5% of Vox voters said they would have voted differently had they known the result, more than double the share among PP or PSOE voters, and only 3.9% were satisfied. Using the 29,000 interviews of the CIS pre-election survey, we also measure where the votes in play were: those who decided during the campaign were concentrated in the audiences of TVE, RNE, El País and the Catalan media, not in those of COPE, TRECE or ABC, which arrived with their vote already decided.

3 November 2026 · Miguel · Leer en español

A shared expectation

The CIS pre-election mega-survey of June 2023 asked each respondent which party they thought would win. 65.5% answered the PP and 19.5% the PSOE. The expectation was not limited to right-wing voters:

90% of PP voters and 82% of Vox voters expected the PP to win. Among PSOE voters, 47% expected the PP to win, more than the share expecting a win for their own party (40%). And more than half of Sumar voters saw the PP as the winner.

The PP did win. But the expectation that went with that win, according to the polls published during the campaign, was of a PP and Vox majority. It did not come: together they won 170 seats (137 and 33), six short of 176.

Afterward: who regretted it

The CIS post-election survey asked whether, had they known the result, they would have voted the same way.

• Vox voters regretted it the most. 13.5% would have voted differently knowing the result, compared with 5.4% of PP voters and 5.2% of PSOE voters. Among Sumar voters, 9.7%.
• And they were the least satisfied. Only 3.9%, compared with 7.5% for the PP, 32.6% for the PSOE and 35.6% for Sumar.

The survey does not ask whom they would have voted for instead. The plausible but unmeasured explanation is that some of those Vox voters would have concentrated their vote on the PP to secure the majority they took for granted. It is the flip side of the tactical vote: they voted for their preferred option because they believed the sum was assured.

Was there a bandwagon effect?

The bandwagon effect assumes that undecided voters join whoever they think will win. Among those who were wavering or had not yet decided in June, the split in who they wanted to win changes a lot depending on who they thought would win:

There is no clear bandwagon toward the favorite. Among undecided voters who saw the PP as the winner, as many wanted the PSOE to win as the PP. What does show up is projection: people who prefer a party tend to believe it will win. With a single survey and no repeated interviews, the two cannot be separated.

Where the votes in play were

Most of the electorate follows the campaign on television and radio. The CIS asked which channel, which station and which newspaper each respondent preferred for following election news. The audiences were highly polarized, and not all arrived equally decided:

• Right-leaning audiences arrived with their vote decided. Among those who preferred COPE, TRECE or ABC, only 15% to 21% decided during the campaign.
• The votes in play were in other audiences. 32% of those who preferred TVE or RNE decided during the campaign, as did 31% of El País readers and between 41% and 49% of those following the campaign on TV3, RAC1 or Catalunya Ràdio. Those audiences held a good part of the vote that, as we explained in “What 2023 taught us,” ended up going to the PSOE and Sumar in the final stretch.
• Mail-in voting, among newspaper readers. 19% of those who preferred El País or El Mundo had requested it or planned to, more than in any television audience, in line with their more highly educated profile.

None of this proves that a media outlet changes anyone’s vote. People who choose a channel or a station already have their leaning, and audiences resemble their voters for that reason. What the data show is where, in June 2023, the electorate that could still move was.

What this means for 29N

• Expectations can demobilize those who think they are winning. In 2023 it was Vox voters who most regretted the result. With the PP and Vox above a majority in almost every 2026 poll, the risk of an expectation taken for granted is again on the same side.
• The vote decided during the campaign is not evenly distributed. If the pattern repeats, the final movement will depend more on the audiences of public television and radio, of the center-left and of Catalonia than on those of the right.

How we calculated it

• Data. CIS microdata: pre-election survey 3411 (June 2023, 29,201 interviews) and post-election survey 3420 (10,101), weighted with their survey weights.
• Expectations. Questions “which party do you think will win the election” and “which would you like to win,” by direct vote intention. Undecided voters: those torn between parties or who have not yet decided.
• Audiences. Preferred television channel, radio station and newspaper “for following political and election news during the campaign.” TVE groups La 1, La 2, the 24 Horas channel and answers with no channel specified; RNE, its stations. Decides during the campaign: “decides during the campaign” or “on the day of the vote.”
• Regret. Post-election survey question on whether they would have voted the same had they known the results, by recalled vote. Recalled vote underrepresents PP and Vox voters, but the percentages are within each electorate.
• Limits. These are correlations from a single survey, not causal effects, and it is not a panel: the same people are not interviewed before and after.
• Code. medios_expectativas_2023.py, in the altnalysis repository.