tag:blogger.com,1999:blog-192445222306631874.post348144202265253922..comments2024-05-23T10:06:33.921-07:00Comments on Xenia Schmalz's blog: Some practical considerations in arguing for the nullXenia Schmalzhttp://www.blogger.com/profile/02238923475669435076noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-192445222306631874.post-45667475132835850012015-10-12T08:54:42.797-07:002015-10-12T08:54:42.797-07:00"Is there a way to change the prior that has ..."Is there a way to change the prior that has less of an effect on the error margins?" Probably not, but I bet you'd be ok with more sensitive data! Maybe ask Morey about it, he'd know more about it than me.<br /><br />I wouldn't worry about the posteriors matching. Cauchy/g priors are designed to represent the information value of very few observations (they are t distributions with ~1 df), so with moderate amounts of data they wash out fast. Models with fat-tailed priors will generally do that. If you're using more informative priors, then naturally they will converge at a rate related to their relative informativeness and the information contained in the data.<br /><br />But in general I don't recommend looking at posteriors to judge support for H0. Too ad hoc and vague. The only principled way to do it is with a Bayes factor :)Alexander Etzhttps://www.blogger.com/profile/07682503927955600974noreply@blogger.comtag:blogger.com,1999:blog-192445222306631874.post-58220835847616646072015-10-08T08:37:53.870-07:002015-10-08T08:37:53.870-07:00Thanks, Alex, that's a great suggestion!
I act...Thanks, Alex, that's a great suggestion!<br />I actually have a question about this - maybe you know the answer or could direct me to a relevant source:<br />I have repeated the Bayes Factor analyses for a study where the default parameters provided evidence for the null, using smaller priors ("rscaleFixed = 0.1"). This has increased the error margin to +/-100%, making the results uninterpretable. Is there a way to change the prior that has less of an effect on the error margins?<br />As an alternative approach, a statistician once told me that one can do the analyses both with a large and with a small prior, and then compare the posteriors of the critical effect: if H0 is true, both values should be about equally close to zero. Would you know of any recommendations for drawing conclusions about whether or not the posterior estimates are similar to each other, or whether this is this something that’s up to the researcher’s judgement?<br />Any advice would be appreciated!Xenia Schmalzhttps://www.blogger.com/profile/02238923475669435076noreply@blogger.comtag:blogger.com,1999:blog-192445222306631874.post-9206504602091548712015-10-07T10:24:05.739-07:002015-10-07T10:24:05.739-07:00Another thing you can do to avoid accusations of c...Another thing you can do to avoid accusations of choosing the "wrong" prior is to re-analyze it yourself and explicitly show that the evidence isn't qualitatively changed by using other reasonable priors. Alexander Etzhttps://www.blogger.com/profile/07682503927955600974noreply@blogger.com