April 22, from 20 hours, even as thousands of voters will just put their ballot in the ballot box, mainstream media announce the results of the first round. But how is this possible? practice, two techniques are used by the media to announce the likely outcome of the election.
operations estimate consist of extrapolating results from the national recounts done in some offices that close at 18h or 19 hours. These offices are selected based on results of previous elections to form a sample of representative offices of the electorate. Then, applying the adjustment models (to account for differences in supply policy with the previous elections), the estimated national results from the first results of these offices. A first estimation takes place at 18:30 (but is not made public to avoid influencing Last voters) and then refined as and when taking into account the additional votes. The first estimate
operation was conducted in 1965 by a team from the AFP, led by political scientist Jean-Luc Parodi, working from the results of a hundred common. She helped to establish as soon as 19 hours h50 General de Gaulle got 42 to 45% of the vote and was therefore tie in the first round (1).
estimation operations generally allow to clearly identify the main trends of the 1st round and give very good results for the 2nd round. But they have two drawbacks. They are very expensive, because they mobilize many people (four to five cents for a presidential election, more so for other elections). They do not provide information on the composition of electorates and the motivations of the vote.
is why the television call, since 1983, a second technique: the exit polls (SSU) . The SSU
are polls on the day of the election with a sample of voters from voting. They are asked to exit their offices on the voting have made and the reasons for their choice.
The main advantage of ESS is to provide data that goes beyond the scores obtained by candidates who can perform an explanation of hot French vote. We can see whether the voters are determined at the last minute or longer before the election, an idea of the direction in which an occupational group or age group voted, or understand how the deferrals are made votes between the two towers. Nevertheless, data from ESS are not very reliable and should be handled with more caution than intention polls to vote. They are based on statements that can not be completely sincere. ESS tend to over-represent voters most politicized and older who respond more readily to requests from investigators and under-represent the people very politicized and having done some studies (2).
________________ (1) But this information was disseminated by the AFP to 9:20 p.m., Chairman of AFP, John Marin (Gaullist history had been one of the voice of Free France broadcast on the BBC) have been extremely cautious before committing.
(2) is certainly can partially correct these biases by various relief methods. But SSU is made extremely quickly, these adjustments are probably even more imperfect they are for traditional political polls.
This post is an excerpt from my book How does one become President (e) of the republic? The strategies of the candidates (Robert Laffont), Chapter 7, where you can find more information on election nights.
0 comments:
Post a Comment