Skip to main content
19 events
when toggle format what by license comment
Apr 7, 2022 at 0:50 history edited kjetil b halvorsen
edited tags
Apr 6, 2022 at 18:49 history became hot network question
Apr 6, 2022 at 18:33 history edited Manuel CC BY-SA 4.0
added 1 character in body
Apr 6, 2022 at 17:47 answer added Tim timeline score: 4
Apr 6, 2022 at 16:58 history edited Manuel CC BY-SA 4.0
added 3 characters in body
Apr 6, 2022 at 16:55 vote accept Manuel
Apr 6, 2022 at 15:00 history tweeted twitter.com/StackStats/status/1511720390929928195
Apr 6, 2022 at 14:29 answer added Sextus Empiricus timeline score: 8
Apr 6, 2022 at 14:11 answer added Christian Hennig timeline score: 8
Apr 6, 2022 at 14:01 answer added Eli timeline score: 4
Apr 6, 2022 at 13:49 comment added Christian Hennig In fact the mean is quite non-robust in a frequentist sense. If the underlying distribution has heavy tails, even with existing variance, the sample size at which the CLT provides a good approximation may be arbitrarily large, i.e. larger than any sample size you may have. (Bayesians should worry about this as well.)
Apr 6, 2022 at 13:31 history edited Manuel CC BY-SA 4.0
deleted 2 characters in body
Apr 6, 2022 at 13:08 history edited Manuel CC BY-SA 4.0
added 683 characters in body
Apr 6, 2022 at 13:04 comment added Manuel @MichaelLew i have added some detail of my intention hope it helps.
Apr 6, 2022 at 13:01 history edited Manuel CC BY-SA 4.0
added 683 characters in body
Apr 5, 2022 at 21:17 comment added Michael Lew Given that there is no information beyond the (large) sample, one would think that even the most Bayesian Bayesian would just calculate the sample mean... What do you really want to know?
Apr 5, 2022 at 21:15 comment added seanv507 i am not sure what your question is. But maybe you are after a bayesian CLT? see stat.columbia.edu/~gelman/book/BDA3.pdf pages 587...
Apr 5, 2022 at 20:53 history edited Manuel CC BY-SA 4.0
deleted 10 characters in body
Apr 5, 2022 at 20:44 history asked Manuel CC BY-SA 4.0