El papel de las encuestas en las elecciones para gubernaturas en México, 2022

Abstract

Apart from what commentators and the lay public usually pay attention to, which is the proximity in the distance estimated by survey and that observed at the polls between the first two places, data that is clearly incomplete except on the rare occasions when it is a duel restricted to two candidates, and the deformations caused by the extraction of the average deviation per contender between what is estimated and what is observed, which is significantly affected by the number of candidates, the provision of two estimators of accuracy corresponding to the Minkowski distance normalized to correspond to percentages, shows that there is a relative persistence in the levels of accuracy achieved by the surveys prior to elections for governorships carried out by personal interviews in homes in 2022 compared to what was recorded a year earlier. This is what would be expected from activity in a professionalized field of activity: few variations in the quality of studies carried out according to methods attached to a scientific canon. Now, it remains to be seen whether this consistency is sustained over a longer period or whether changes occur that could derive from factors endogenous to research work, such as improvements in methods and techniques, or exogenous factors, such as changes in the conditions of political competition.