tag:blogger.com,1999:blog-5880610.post4141257579131318619..comments2024-01-13T18:57:18.243-05:00Comments on Information Processing: Replication is hard; understanding what that means is even harderSteve Hsuhttp://www.blogger.com/profile/02428333897272913660noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-5880610.post-88476643085610416582015-05-06T17:52:17.810-04:002015-05-06T17:52:17.810-04:00;-)
I don't know for sure -- let's see wh...;-)<br /><br />I don't know for sure -- let's see what the Science paper says when it comes out.steve hsuhttp://plus.google.com/u/0/+StephenHsu/aboutnoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-14805295494581431902015-05-06T17:49:18.537-04:002015-05-06T17:49:18.537-04:00What does it even mean to say that the non-replica...What does it even mean to say that the non-replicating experiment had "similar results" to the original? Could it be something as simple as that the overall "effect" is in the same direction as in the original? For many designs, wouldn't that happen half the time anyway?outside_observernoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-41930412265269421432015-05-04T17:15:15.674-04:002015-05-04T17:15:15.674-04:00Always a good read:
http://www.economist.com/news/...Always a good read:<br />http://www.economist.com/news/briefing/21588057-scientists-think-science-self-correcting-alarming-degree-it-not-trouble/RealityIsComplicatednoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-88119628560392690902015-05-04T15:59:43.012-04:002015-05-04T15:59:43.012-04:00I would recommend reading the article by David Col...I would recommend reading the article by David Colquhoun that he linked in the comments of the news article: http://rsos.royalsocietypublishing.org/content/1/3/140216 <br />It's rough contention is that total lack of understanding of what p values mean gives a false discovery rate of 30% in the best case scenario.<br />As for the claim that 24 of the studies were almost significant are they Mark Woodleynoreply@blogger.comtag:blogger.com,1999:blog-5880610.post-87289171390570384082015-05-04T10:37:32.446-04:002015-05-04T10:37:32.446-04:00" they typically have not actually grasped th..." they typically <i>have not actually grasped the implications for their own work</i>"<br /><br />Most of them don't know how to use Bayes theorem even in the simplest case, such as the sensitivity and specificity of a test. Most of them often confuse significance with effect size. Would you please spell out the implications in a post?MUltannoreply@blogger.com