The statistical analysis of vote transfers indicates that the main beneficiary of the greater attendance of voters last Sunday with respect to the primary elections – 78% against 70%, an increase well above the average of those registered among these instances – was Sergio Massa: won the preferences of at least half of the voters who were absent in August. These citizens also supported Javier Milei and Patricia Bullrich but to a much lesser extent. The Minister of Economy also appears favored by the bulk of those who voted for pre-candidates who in August did not exceed the legal minimum floor.
Another notable transfer is the more than two points that voted for Milei in primaries and that now appear preferring Juan Schiaretti, although this estimate is close to the limit of the confidence level of the applied method. The analysis also shows that during the first round the candidates preserved the overwhelming majority of the votes to their groups in primaries, in varied proportions. Also notable is the greater transfer of blank votes of the primaries towards Massa than towards his competitors, although again at the limit of the level of trust.
The methodology
The diagrams that illustrate this note show vote transfers. On the left are the percentages of votes for each electoral option in primaries and on the right are the percentages in the first round, both calculated on all the voters in either of the two instances, which includes those absent in one who voted in the other. The lines that join them and their thickness represent the percentage points that we estimate were transferred. All analysis percentages include blank votes, null votes and absentees who voted in the other instance. Hence, they do not coincide with the percentages of the official results, calculated on valid votes in primaries (that is, votes for groups plus blank votes) and on affirmative votes (only for groups) in the first round.
To estimate these voter flows, we apply a statistical model called «ecological inference» developed by political scientist Gary King (Harvard University) that has been accepted in United States courts to settle cases of racial discrimination in the demarcation of electoral districts. The results obtained have been essentially the same using the establishments or the tables. The results database comes from the data service made available to the media by the National Electoral Directorate and part of the code for the implementation of the model was provided by the political scientist Ernesto Calvo from the University of Maryland. Additionally, we have an application that allows you to view the transfer estimation diagrams in each province and the Federal Capital at the table level.
It is also important to note that small values (say, below two percent) can contain considerable margins of error and those fine lines that cross the diagram appear only because the model imputes the fate of all votes. This means in this case that the estimates that should be interpreted with more caution are those flows of smaller amounts and those that involve the least voted electoral options. Furthermore, it is always advisable to contrast these estimates with those obtained through surveys and other opinion measurements.
October Group Data Team
* Political scientist
**Computer scientist
Esta nota es parte de la red de Wepolis y fué publicada por Jhon Williams el 2023-10-25 16:05:04 en:
Link a la nota original
Palabras clave:
#greater #participation #key #rebound #Sergio #Massa
