How are the surveys conducted? Trial and defense

Today I defend the usefulness of opinion polls after a volatile election night. In fact, I did the exercise of writing the gist of this text on a Sunday afternoon, before knowing the results. It is indeed a habit that helps me think without defeatism or demographic euphoria: I will prove the usefulness of opinion polls anyway.

How did they do it? normal.

The average guessed the order of the large parties and the distance between the two traditional parties and their pursuers. The average error was 1.2 points, much better than the historical average (1.9). But he grossly overestimated the right (who wins by one point, not six) and fell short with the PSOE. In favor of the polls, it must be said that they left the result very open-ended – at least as a whole. for me latest predictionBased on the average and historical accuracy of good polls, it gave a 60% chance to the majority of the right; 16% for a left-wing success allowing him to govern with his support from 2019; And 23% that neither one nor the other happened, and that we got into a situation that almost broke down. Reality chose that middle ground.

This was the most likely scenario according to 40dB scan. It was published by this newspaper last Monday. His poll was quite accurate: it gave 173 seats to the combined PP and Vox, which stayed at around 169. The biggest mistakes were made by the pollsters, whose sum exceeded 180. The discrepancies between one and the other somehow facilitated my work, because it is easier to explain uncertainty when there is a discrepancy in the criteria. Other times it wasn’t. In his most remembered refereeing of the decade, when he was in 2016 a surprise From Podemos to PSOE, it was hard that all the surveys gave it as a potential. It wouldn’t matter if it weren’t for the huge difference in votes. This consensus of polling firms conveys a false sense of certainty.

The CIS deserves separate attention.

The left has fared much better than expected in the rest of the polls, but nonetheless the CIS remained biased. His estimate was as accurate as average — averaging 1.16 points per game, up from 1.2. But his deviation goes in the foreseeable direction: He once again exaggerated the total strength of the left. Since José Félix Tezanos came to the CIS leadership, the center has overestimated votes from the left in 36 of 37 elections. You must validate your form He hits Six times in a row.

What affects most is what happens closer. In order not to miss anything, subscribe.

Participate

The value of surveys

Why do I demand opinion polls? First, because they provide information. Last week teaches us what happens when rigorous polling is turned off: the alternative is not reflective and peaceful silence, but noise, shenanigans and intoxications of self-interest. Without an election, all parties will say they win. Everyone may claim they have a seat at stake in your county; Every vote will be helpful.

Opinion polls were promoted by the American media to provide data that acted as a counter force to politicians. Transparent data — a commitment we doubled down on in this paper this cycle — is independent and rigorous. This task remains with us.

Secondly, because opinion polls make elections more predictable. Thanks to them we know what is possible, probable or very rare; It is legitimate for some citizens to use that information for business. To cast a useful vote or to mobilize and stop a party they do not want to see in the institutions. In the end, the entire democratic system can be thought of as a process: millions of different people reconciling our differences until we agree on certain government and laws. Good surveys have a role in this process.

third reason? Readers want to take polls, we work with them and we owe it to them.