[R] ANOVA "Types" and Regression models: the same?

Frank E Harrell Jr f.harrell at Vanderbilt.Edu
Tue Mar 2 15:01:18 CET 2010


Sorry there were 2 typos in my note:
> 
> John Fox's posting provided a lot of information.  Briefly, the "Types" 
> refer to whether effects are adjusted for all other effects in the model 
> (Types II, III, IV) or no (Type I or sequential tests only adjust for 

no -> not

> EARLIER terms in the model).  See John's posting for definitions of 
> Types II and III and for reasons to almost always use II or III. 

II or III -> II over III

Sorry about that
Frank



More information about the R-help mailing list