Comments on: There are no shortcuts to knowledge http://blogs.discovermagazine.com/gnxp/2012/11/there-are-no-shortcuts-to-knowledge/ Sun, 19 May 2013 01:40:00 +0000 hourly 1 http://wordpress.org/?v=3.4.2 By: Charles Nydorf http://blogs.discovermagazine.com/gnxp/2012/11/there-are-no-shortcuts-to-knowledge/#comment-48661 Charles Nydorf Mon, 26 Nov 2012 13:50:35 +0000 http://blogs.discovermagazine.com/gnxp/?p=19159#comment-48661 When I was growing up in the 1960's, the climate of opinion was strongly frequentist. The subsequent swing towards Bayesianism is one of the most fascinating intellectual developments that I have lived through. When I was growing up in the 1960′s, the climate of opinion was strongly frequentist. The subsequent swing towards Bayesianism is one of the most fascinating intellectual developments that I have lived through.

]]> By: Jim http://blogs.discovermagazine.com/gnxp/2012/11/there-are-no-shortcuts-to-knowledge/#comment-48660 Jim Mon, 26 Nov 2012 13:23:30 +0000 http://blogs.discovermagazine.com/gnxp/?p=19159#comment-48660 I agree that confirmation bias (CB) should be avoided, but your method, accept hypothesis A if the most popular (or even most plausible) criticism of A seems false (or even is false), has obvious flaws. Given that you never intended to quantitatively estimate election outcome probabilities and that CB was a significant concern, I wonder why you did not entirely eliminate potential CB by constructing your bets so that amount of your winnings did not depend on the outcome. (This approach has the obvious drawback that you cannot maximize your likely winnings based on your estimate of the probability of an Obama win, but you did not seem to take that approach anyway.) I agree that confirmation bias (CB) should be avoided, but your method, accept hypothesis A if the most popular (or even most plausible) criticism of A seems false (or even is false), has obvious flaws. Given that you never intended to quantitatively estimate election outcome probabilities and that CB was a significant concern, I wonder why you did not entirely eliminate potential CB by constructing your bets so that amount of your winnings did not depend on the outcome. (This approach has the obvious drawback that you cannot maximize your likely winnings based on your estimate of the probability of an Obama win, but you did not seem to take that approach anyway.)

]]> By: Thomas Ferraro http://blogs.discovermagazine.com/gnxp/2012/11/there-are-no-shortcuts-to-knowledge/#comment-48659 Thomas Ferraro Mon, 26 Nov 2012 12:40:24 +0000 http://blogs.discovermagazine.com/gnxp/?p=19159#comment-48659 The problem with the argument of those who claimed (pre-election) that the polls were skewed because turnout was going to be different than 2008 is that polls account for that by asking respondents if they are going to vote. And the turnout was really their only argument that wasn't ad hominem. The poll aggregators (Silver, et al), weighed likely-voter polls more heavily than registered-voter polls. And almost all the polls report only likely-voter results as election gets closer. Those who argued polls were skewed were really arguing poll respondents were lying about whether they were going to vote. As I recall, Silver addressed this in one or more of his posts. The problem with the argument of those who claimed (pre-election) that the polls were skewed because turnout was going to be different than 2008 is that polls account for that by asking respondents if they are going to vote. And the turnout was really their only argument that wasn’t ad hominem.

The poll aggregators (Silver, et al), weighed likely-voter polls more heavily than registered-voter polls. And almost all the polls report only likely-voter results as election gets closer.

Those who argued polls were skewed were really arguing poll respondents were lying about whether they were going to vote. As I recall, Silver addressed this in one or more of his posts.

]]>