3 ms·

I'm still waiting for somebody to talk about how tightly coupled Nate Silver's forecasts are to the political processes he attempts to forecast. What if Nate S

I'm still waiting for somebody to talk about how tightly coupled Nate Silver's forecasts are to the political processes he attempts to forecast. What if Nate Silver were to run around on the football field before each play whispering in the players' ears about the how likely they are to win, according to his latest forecast. Imagine that they believed in him, or at least felt some degree of superstition when he came around. I just think there's something innately flawed about public forecasting of an election. Mathematically and maybe even ethically. I'm still trying to figure out how to articulate why I feel that way. It could even be argued that polling itself could influence opinions by introducing bias at a critical moment when someone is being asked to consider who they are voting for. What if that's the moment when they make up their mind? P.S. https://www.frontiersin.org/articles/10.3389/fphy.2015.00077/full https://www.frontiersin.org/articles/10.3389/fphy.2015.00077...

dagw 8y ago
I just think there's something innately flawed about public forecasting of an election. It used to be the case in France that you weren't allowed to publish polls two weeks before an election for just this reason. They eventually scrapped this with the argument that the all 'elites' had access to internal and private polls anyway so all you where doing was denying the 'people' the same access to information that the 'elites' had. It could even be argued that polling itself could influence opinions by introducing bias at a critical moment when someone is being asked to consider who they are voting for. This is a pretty well known effect, to the point that many political strategists try to avoid publishing or drawing attention to polls showing their candidate too far ahead in case it makes voters complacent and decide to stay home. The ideal situation is to make people convinced the polls show the candidates tied 50-50 going in to the election and that 'your' vote could easily swing the whole election.
- mijamo 8y ago
  This is slightly wrong. It was initially one week, not two. It was changed because it was unconstitutional, as it limited freedom of expression too much, not because it was useless or anything to do with elite vs people. And it still exists, limited to the day of the election and the day before. Source: https://www.conseil-constitutionnel.fr/nouveaux-cahiers-du-conseil-constitutionnel/la-reglementation-des-sondages-et-l-election-presidentielle https://www.conseil-constitutionnel.fr/nouveaux-cahiers-du-c...

dagw 8y ago

I just think there's something innately flawed about public forecasting of an election. It used to be the case in France that you weren't allowed to publish polls two weeks before an election for just this reason. They eventually scrapped this with the argument that the all 'elites' had access to internal and private polls anyway so all you where doing was denying the 'people' the same access to information that the 'elites' had. It could even be argued that polling itself could influence opinions by introducing bias at a critical moment when someone is being asked to consider who they are voting for. This is a pretty well known effect, to the point that many political strategists try to avoid publishing or drawing attention to polls showing their candidate too far ahead in case it makes voters complacent and decide to stay home. The ideal situation is to make people convinced the polls show the candidates tied 50-50 going in to the election and that 'your' vote could easily swing the whole election.

mijamo 8y ago
This is slightly wrong. It was initially one week, not two. It was changed because it was unconstitutional, as it limited freedom of expression too much, not because it was useless or anything to do with elite vs people. And it still exists, limited to the day of the election and the day before. Source: https://www.conseil-constitutionnel.fr/nouveaux-cahiers-du-conseil-constitutionnel/la-reglementation-des-sondages-et-l-election-presidentielle https://www.conseil-constitutionnel.fr/nouveaux-cahiers-du-c...