Talk:Explained sum of squares

Latest comment: 13 years ago by Melcombe in topic Terminology

isn't there a a sample mean missing from the equation?— Preceding unsigned comment added by 64.229.104.106 (talk) 00:43, 25 February 2006

Yes, there was, and I've fixed it. Can you believe it was also missing from the Total sum of squares article. Btyner 04:20, 10 March 2006 (UTC)Reply

the grand mean

edit

how is it defined? It is the average of all y. Jackzhp 22:03, 12 October 2006 (UTC)Reply

Terminology

edit

Anyone know a reference that uses "Type I SS" and "Type III SS" as in present article? Melcombe (talk) 15:03, 18 August 2008 (UTC)Reply

I do... whoever wrote this seems to be thinking in terms of a software package SAS. See link that follows. It isn't clear to my why in the article for ESS we're explaining model specific versions of different types of SS. After looking for an hour of so it still isn't clear to me that using these other types will give something analogous to ESS. If it is decided to keep this section, there is a forth "type" that isn't mentioned in our wikipedia article that should probably be mentioned just to be complete.
http://www.psych.umn.edu/faculty/waller/classes/mult10/readings/sasSOS.pdf —Preceding unsigned comment added by 160.94.156.170 (talk) 14:38, 18 March 2010 (UTC)Reply
I agree -- the sections on Type I SS etc. contradict the rest of the article and don't belong here, so I'm deleting them. Duoduoduo (talk) 20:13, 18 December 2010 (UTC)Reply

I seem to recall having heard of four types, but I really only learned one of them, and I remember what its SAS Roman numeral was, but it's the one conventionally reported in sequential ANOVA table. Michael Hardy (talk) 20:18, 18 December 2010 (UTC)Reply

:These recents edits are moving towards duplication of what is in both Sum of squares and Squared deviations. So a merging may be useful. Melcombe (talk) 10:18, 21 December 2010 (UTC)Reply

Avoiding Confusion

edit

I have found that in many cases the ESS is refered to as the Sum of Squares due to Regression. This is confusingly also know as SSR, but is not the same as Sum Square Residuals (which is also known as the Sum of Square Errors, SSE). I believe that it should be changed to reflect this. Furthermore the disambiguation page for SSR should point to both types of SSR so as to not lead to confusion. RobJN (talk) 13:45, 19 October 2008 (UTC)Reply

Expert tag

edit

I added the expert tag because of the above unanswered questions, and because of the lack of citations. Melcombe (talk) 14:07, 18 September 2009 (UTC)Reply