Mail Archive Search: <+subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*model\s+performance\s+with\s+and\s+without\s+interactions\.\.\.\s*$/>

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*model\s+performance\s+with\s+and\s+without\s+interactions\.\.\.\s*$/: 3 ]

Total 3 documents matching your query.

1. Re: model performance with and without interactions... (score: 1): Author: "Fowler, Mark" <FowlerM@mar.dfo-mpo.gc.ca>; Date: Fri, 20 Jun 2003 11:53:27 -0300; Yes, we did that. That was exactly the justification for removing interaction terms (the interactions did not suggest troubling differences in trends, just differences like greater and lesser slopes; /archives/html/s-news/2003-06/msg00147.html (12,055 bytes)
2. Re: model performance with and without interactions... (score: 1): Author: Naomi Altman <naomi@stat.psu.edu>; Date: Fri, 20 Jun 2003 11:55:36 -0400; I would not drop interactions based on the relative size of F-tests. I look at the interaction plots, as suggested here. Some statisticians would NEVER drop interactions, because there are inferentia; /archives/html/s-news/2003-06/msg00148.html (13,178 bytes)
3. Re: model performance with and without interactions... (score: 1): Author: Spencer Graves <spencer.graves@PDF.COM>; Date: Fri, 20 Jun 2003 09:30:25 -0700; George Box (1999) has said that the proper role of statistics is to catalyze the process of scientific learning. What one does with the interactions depends at least partly on whether one is doing ex; /archives/html/s-news/2003-06/msg00149.html (16,216 bytes)

Current List: 1 - 3
Page: [1]