Mood waves and electoral polls

Something came to my mind that I have already witnessed in some electoral campaigns in which I have participated in recent years. Companies specialized in demoscopy are frequently approached to carry out “tailor-made” electoral surveys, as one more campaign tool, in order to serve the objective of mobilizing or demobilizing voters.. I have never seen behavior that was not professional, but it is evident that the proximity to a certain candidacy introduces a voluntaristic bias —especially when the results are very close— that could condition, in a certain way, the result.

The failure of the surveys is not in the size of the samples (the CIS has handled samples of more than 19,000 valid interviews), nor in the methodology (the one that came closest, 40DB for PRISA, used online surveys).. The reason is, from my point of view, in the way in which the parties make us interpret the numbers, especially in internal polls and in which they even estimated 160 seats for the Popular Party.

In the words of the sociologist and professor Juan José Toharia, one of the most authoritative voices in the Spanish desmoscopy, “leaving aside the CIS, which plays in another league and is not right even when it is wrong, we have to further refine our ability to capture emotional waves , which for now seem to be below what our radar can measure with due precision”.

I am not an expert in demoscopy. I collaborated with the Albert Rivera survey team in 2015. The work consisted of “capturing the psychic waves” to which Toharia referred, to clarify the existing surveys and thus achieve a more accurate picture.. I reviewed the work this morning and the truth is that the electoral forecasts made in October of that year “nailed” the December electoral results thanks to introducing variables more linked to reputation management into the equation. It is true that a success does not validate a model.

The main conclusion I got from that experience is that in traditional polls positive factors tend to prevail over negative ones: sentiment, memory, voting intention, trends, etc.. All of them are positive factors that, in an ideal world, could translate into voting. But the reality is that there are barriers and difficulties that prevent this intention from finally coming to fruition.. If the resistance to voting is not well calibrated, the danger is precisely that inertia (the “wave” some say) leads to maximizing positive data. This is precisely what happened to the PP on 23-J.

But what are the psychic waves that should be considered to refine more in the future? The reputational model to which I have referred before, and I quote professors Carreras and Alloza, considers emotion as a determining factor of the intention (to vote). And emotion is built through two variables that are self-identity and social identity, which, in turn, takes into account factors such as social image, moral pressure, social pressure, convenience, and the perception of ease or difficulty of change.

All these factors have been put into play in this electoral campaign. The image, the moral pressure and the social pressure have been present in a purely ideological campaign between blocks (left and right) where the PP pacts with Vox have undoubtedly retracted part of the more moderate voters who could opt for the PP as political change option. The voluntarist bias led to underestimating the political cost of these pacts, because in the PP they felt that they were facing a “wave” of change, as one of the members of the party leadership explained to me before the elections.. And this perception was contagious in the results of the surveys.

These factors were predictable, although they were underestimated.. The convenience factor was also foreseeable, which was highly commented. In this case, the convenience was conditioned by the date of the electoral call in the middle of the summer season. Under normal conditions, the day chosen for the elections should have resulted in a low turnout. The same as happens when the electoral appointment occurs on a particularly rainy day. Low turnout would have confirmed the wave of change. It was known that the high turnout —in the end it was 70.20%, higher than that of 2019— would favor the electoral expectations of President Pedro Sánchez and the PSOE, because it was the option that had the most to lose. The high number of applications to vote by mail could have served as an alert to qualify the PP results downwards, but it was not possible to analyze them correctly either.

However, the factor (mood wave) that has most influenced the forecasts, in my opinion, is the “perception of ease or difficulty” of the change. Although some voices recalled the “capacity for resistance” of Pedro Sánchez, the PP campaign started from an excess of confidence due to the good results of the Autonomous and Municipal elections of 28M. A confidence that underestimated the chances of a comeback for the PSOE. Despite the fact that the President of the Government risked a very personal campaign to which he gave himself body and soul, raising the morale of his people after the crash of 28M and demonstrating that his commitment and conviction was total —to the point of placing the campaign over a project as desired as the Spanish Presidency of the EU—, Feijóo's victory in the face-to-face debate for many analysts was confirmation that Pedro Sánchez was finished.

The failure of the 23J electoral polls highlights the need to refine the ability to capture the “mood waves” of voters. Surveys, while useful for capturing general trends, should more carefully consider negative factors that may influence voting intentions, such as resistance to voting.. Closeness to certain candidates and the influence of public poll results can skew forecasts and lead to underestimating important political costs.. In the future, it will be crucial to incorporate emotion and factors such as self-identity and social identity into the analysis, which can condition the perception of ease or difficulty in the face of change.. This is the only way to improve the accuracy and reliability of electoral polls in an increasingly complex and dynamic political context..

*Juan Cardona Soriano is a director of PROA Comunicación