[R] GLMM and data manipulation (2nd try)

[Giovanni Petris]

Andrew,

Thank you for your reply. In fact, I had a way of unrolling the table
but I think yours look much nicer - much readable to me. Below is what
I did, but I was afraid I would scare my students away from R!

> approval <- factor(c("Approve", "Disapprove"),
+                    levels = c("Disapprove", "Approve"))
> survey <- factor(c("First", "Second"))
> tmp <- data.frame(approval = unlist(expand.grid(approval, approval)),
+                   survey = rep(survey, each = 4))
> rat.df <- cbind(tmp[rep(1:8, rep(rating, 2)), ],
+                 id = factor(rep(1:sum(rating), 2)))
> row.names(rat.df) <- NULL

Thank you again,
Giovanni

> Date: Sat, 03 May 2008 06:43:22 +1000
> From: Andrew Robinson <A.Robinson at ms.unimelb.edu.au>
> Cc: r-help at r-project.org
> User-Agent: Mutt/1.4.2.3i
> 
> On Fri, May 02, 2008 at 03:06:31PM -0500, Giovanni Petris wrote:
> > 
> > Hello,
> > 
> > I posted a question yesterday but I got no replies, so I'll try to
> > reformulate it in a more concise way. 
> > 
> > I have the following data, summarizing approval ratings on two
> > different surveys for a random sample of 1600 individuals:
> > 
> > > ## Example: Ratings of prime minister (Agresti, Table 12.1, p.494)
> > > rating <- matrix(c(794, 86, 150, 570), 2, 2)
> > > dimnames(rating) <- list(First = c("approve", "disapprove"),
> > +                          Second = c("approve", "disapprove"))
> > > rating
> >             Second
> > First        approve disapprove
> >   approve        794        150
> >   disapprove      86        570
> > 
> > I would like to fit a logit model with approve/disapprove as response,
> > survey (first/second) as a fixed effect, and subject as a random
> > effect. 
> > 
> > 1) Is it possible to fit such a model directly using "lmer"?
> > 
> > or 
> > 
> > 2) Should I unroll the table above into a dataframe containing also
> >    fictitious subject id's? If this is the case, what is a clean way
> >    of doing it? 
> 
> 
> Unroll it. 
> 
> Asking for a "clean" way to do something is a disincentive because it
> implies that you know how to do it but not cleanly.  In the future I
> would suggest that you do one of two things
> 
> a) post your on dirty version and ask for a cleaner one, or
> 
> b) just ask for something that works.
>  
> Something like ...
> 
> ratings <- 
> 	data.frame(
> 	response = c(
> 		 rep(c(1,1), 794), 
> 		 rep(c(1,0), 150),
> 		 rep(c(0,1),  86),
> 		 rep(c(0,0), 570)),
> 	time = rep(c(1,2), 1600),
> 	subject=rep(1:1600, each=2))
> 
> test.lmer <- lmer(response ~ time + (1|subject), data=ratings,
>      	     family=binomial)
> 
> but I don't know if you think that's clean or not.
> 
> Andrew
>  
> > Thank you in advance,
> > Giovanni Petris
> > 
> > -- 
> > 
> > Giovanni Petris  <GPetris at uark.edu>
> > Associate Professor
> > Department of Mathematical Sciences
> > University of Arkansas - Fayetteville, AR 72701
> > Ph: (479) 575-6324, 575-8630 (fax)
> > http://definetti.uark.edu/~gpetris/
> > 
> > ______________________________________________
> > 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.
> 
> -- 
> Andrew Robinson  
> Department of Mathematics and Statistics            Tel: +61-3-8344-6410
> University of Melbourne, VIC 3010 Australia         Fax: +61-3-8344-4599
> http://www.ms.unimelb.edu.au/~andrewpr
> http://blogs.mbs.edu/fishing-in-the-bay/ 
>