[R] Creating a model with fixed and random variables

[William Dunlap]

> %in% doesn't generally mean 'nested in' in R. It is a set membership test

In a formula (involving a tilde) given to lm() or glm() %in% generally does mean nesting.
  > attr(terms(y ~ (x1+x2) %in% (x3+x4+x5)), "term.labels")
  [1] "x1:x3:x4:x5" "x2:x3:x4:x5"
but the "|" operator stops terms() from interpreting the usual formula operators
in the usual way
  > attr(terms(y ~ 1 | (x1+x2) %in% (x3+x4+x5)), "term.labels")
  [1] "1 | (x1 + x2) %in% (x3 + x4 + x5)"

Bill Dunlap
Spotfire, TIBCO Software
wdunlap tibco.com


> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On Behalf
> Of S Ellison
> Sent: Monday, August 05, 2013 4:55 AM
> To: cbarl3y; r-help at r-project.org
> Subject: Re: [R] Creating a model with fixed and random variables
> 
> > Code I have used thus far without being able to replicate the
> > data includes:
> >
> > Fm<-lmer(Score~(1|Line%in%Set)+Set+(1|Block))
> >  (I figured out how to get a p-value, but it didn't yield the
> > same results as those obtained in SAS)
> 
> %in% doesn't generally mean 'nested in' in R. It is a set membership test and will return
> TRUE for those labels in Line that are also in Set and FALSE otherwise.
> 
> Did you mean Score~(1|Set/Line)...?
> 
> If you did, bear in mind that , combined with the fixed Set term, (1|Set/Line) implies a
> random Set grouping effect as well as a fixed effect - not sure that makes sense in your
> circumstance unless Set is a continuous predictor. May be safer to define SetLine<-
> interaction(Set, Line) and do
> Score~Set + (1|SetLine) + (1|Block)
> 
> S Ellison
> 
> *******************************************************************
> This email and any attachments are confidential. Any use...{{dropped:8}}
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.