Casillas Poll: Vox Hits 31.6% in Non-Representative Survey

Casillas poll with ten times more votes than CIS puts Vox at 31.6%

The Iker Casillas poll on the general elections gathers ten times more votes than the CIS and yields a distribution no polling firm would endorse as is: PP 36%, Vox 31.6%, PSOE 26.1%, and Sumar 6.2%. Voting and viewing results require registration. The participation volume is staggering; representativeness is debatable. That is the whole issue.

What does the Iker Casillas poll measure and compare to?

The first argument is about size. The CIS, which conducts the most interviews, doesn't reach 30,000; here they claim ten times more. Selling that volume as methodological superiority is the classic error of looking at the sample while forgetting who it represents. A huge sample of a specific audience remains a specific audience. Period.

The second argument concerns alignment with other polls. The PP moves within the margin granted by polls; the PSOE likewise; Sumar appears in half its range. Vox is the only one breaking all forecasts, with more than double the margin conventional polls concede. When a single party breaks the pattern, suspicion falls on the pattern or the respondents.

Why social media polls aren't representative samples

The census of a vote spread on social media resembles the electoral census little. It points to overrepresentation of men, middle-aged individuals, and urban origins, leaving part of the country out irreparably: pensioners barely use these networks, nor do many middle-aged people. This isn't a slight bias. It's half of Spain missing from the count.

The mechanics don't help either. Only those who register vote, and registrants are usually eager supporters, not lukewarm voters. On top of this filter lies the standard doubt in open polls: how many votes are automated. There's a subtle detail often missed by those focusing only on percentages: Casillas' question isn't literally "Who will you vote for?" but a different one, completely changing the result's interpretation.

Add the specific audience profile. The most transversal electorates, PP and PSOE, appear with percentages nearly identical to any poll; the niche parties, Sumar and Vox, sink or surge depending on where the bias blows. No statistics master's degree is needed to see the direction.

Vox's hidden vote: from stigma to 18 points

The most repeated hypothesis is the shame vote. With a stigmatized party, voters don't confess their choice over the phone, so polls fall short; in an open poll among sympathizers, that same voter responds without issue. From there to projecting Vox around 18% is a considerable leap. Not impossible. But not data.

Against the hidden vote thesis weighs history. Four years ago Vox obtained 3,400,000 votes, and those lowering expectations now argue much of that floor came with Ciudadanos' collapse and rejection of Casado, not consolidated loyalty. Some predict the party should be happy retaining thirty MPs. The regional surprise five years ago proves polls fail; conversely, internet polls fail more and almost always in the same direction.

Pensions, postal voting, and campaign noise

Parallel to percentages circulate data on pension revaluation contradicting campaign discourse: 1% hikes with CPI at 2.4% (2012), 1% with 2.9% (2013), and 0.25% with 1.5% (2017). These figures are blamed on the PP from the opposite side, reminding us that underlying economic debate doesn't appear in intent-to-vote polls. Campaign headlines and historical series rarely coincide.

Suspicions about postal voting and counting also surface. To be clear: these are suspicions without proof, the same type seen in every previous call that never amounted to anything. Mutual distrust is already a structural antiestéticature of the system, not new to this campaign.

What to expect? My bet, with all reservations: if hidden vote exists, Vox will exceed poll predictions but remain far from the 31.6% granted by Casillas' poll. If not, it becomes another example of giant sample with tiny audience.