tag:blogger.com,1999:blog-3112258799568696095.post2197233739672277624..comments2024-03-06T13:50:29.718+05:30Comments on E's flat, ah's flat too: Odds are, it's wrongRahul Siddharthanhttp://www.blogger.com/profile/04809667965184094636noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-3112258799568696095.post-68588867783170213012010-03-26T12:15:57.679+05:302010-03-26T12:15:57.679+05:30Kapil -- I am currently reading Jaynes' "...Kapil -- I am currently reading Jaynes' "Probability theory: the logic of science", where he basically trashes "orthodox statistics" (it's an unsystematic collection of ad-hockisms, a hodgepodge of the feuding schools of Fisher and Neyman-Pearson, according to him) and argues that physicists and astronomers have been better off not using statistical hypothesis tests because their own intuition is much better at assessing significance. He argues strongly for Bayesian methods as outlined by him; the methods go back decades and he gives huge credit to Harold Jeffreys and his 1939 book on probability and statistics. <br /><br />Today the worst of the polemics between the "frequentist" and "Bayesian" camps are over and Bayesian methods are widely accepted because they work, but I suspect the deeper points that Jaynes makes are still not widely accepted, perhaps because so many people would be personally hurt by it. For example, he says the chi-square test is often wrong by orders of magnitude because it makes incorrect assumptions about the tails of normally-distributed data. There are several other examples in each chapter.<br /><br />I am very far from finishing the book, much less absorbing it, but will blog about it one of these days.<br /><br />Jai -- yes, that case seemed extremely odd, and like you, it seemed to me that the lady was being victimised.Rahul Siddharthanhttps://www.blogger.com/profile/04809667965184094636noreply@blogger.comtag:blogger.com,1999:blog-3112258799568696095.post-85026170570779253712010-03-23T19:38:04.643+05:302010-03-23T19:38:04.643+05:30I remember reading up a bit on Munchausen by proxy...I remember reading up a bit on Munchausen by proxy a long while ago when there was a spam flood for "rescue Aditya Chandran" on the web and from that site and the best of my memory it sounded a bit iffy by itself; and especially doubtful applied in that case which sounded like a bad divorce going hellishly wrong for the lady concerned.<br /><br />thanks,<br />JaiAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-3112258799568696095.post-61671297394261741552010-03-19T07:11:27.116+05:302010-03-19T07:11:27.116+05:30You will be pleased to note that a basic course on...You will be pleased to note that a basic course on Probability and Statistics is part of the Core that is done by all students at IISER.<br /><br />This at least introduces them to the possibility of methods affecting the outcome of a measurement (Bertrand's paradox). Whether they will recognise it when they see it during their research is another story; after all even statisticians and probability theorists can sometimes be fooled!Kapil Paranjapehttps://www.blogger.com/profile/00136865779726911943noreply@blogger.com