Polls for Governorships in Mexico Accuracy According to Approach Mode (2021-2023)

Abstract

This text reviews the accuracy of the estimates resulting from polls published close to the elections for governors in Mexico held from 2021 to 2023 according to different methods of approximation currently available.

The first part of the essay is dedicated to exposing the characteristics of the indicators conventionally most used to estimate the accuracy of quantitative demoscopic studies aimed at measuring electoral preferences prior to elections, as well as indicators that could be more correct and efficient from a theoretical-methodological point of view for measuring the phenomenon.

Once the traditional estimators and those that methodologically could be more correct and efficient to measure the accuracy of the surveys have been exposed, a comparison is carried out between what was foreseen by the exercises made public in the reference period against the electoral results, differentiating the data according to the approach mode used.

The main limitations of the exercise is its casuistic nature, so the conclusions reached can only be considered as circumstantial evidence and in no way definitive about the phenomenon under study. Its originality, however, lies in the fact that the comparison of differences according to the method of approximation is not usually the way used to analyze the accuracy of the surveys, despite its eventual relevance.

This is confirmed by observing that, as a result of the exercise to estimate the accuracy of the surveys, it is found that this accuracy does depend on the approach method used to carry it out.